Ancient Greek Shorthand.

Candida Moss has an extraordinarily interesting Daily Beast piece on something I hadn’t been aware of:

Several years ago, Ryan Baumann, a digital humanities developer at Duke University, was leafing through an early 20th-century collection of ancient Greek manuscripts when he ran across an intriguing comment. The author noted that there was an undeciphered form of shorthand in the margins of a piece of papyrus and added a hopeful note that perhaps future scholars might be able to read it. The casual aside set Baumann off on a new journey to unlock the secrets of an ancient code.

Initially, Baumann told me, he thought that perhaps everything had been deciphered. “I thought to myself, ‘Well, it’s been about 100 years, maybe someone has figured it out!’ So, I looked into it, and to my delight, the system of ancient Greek shorthand does seem to have been largely figured out.” To his dismay, though, this century-spanning scholarly achievement has also been largely overlooked and underexplored. Very few people are interested in shorthand.

Why does this matter? Well, ancient Greek and Latin shorthand (also known as stenography or tachygraphy) were the bedrock of ancient writing and record keeping. The scripts that emerged in the first century BCE allowed people to record things faster than “normal.” Just like today, said Baumann, stenography was “crucially important” for recording courtroom proceedings and political speeches, but dictation was also used to compose letters, philosophy, and narrative. Everything from ancient romance novels to foundational political theories were first transcribed in shorthand. Often this would have happened on erasable wax tablets (we have many examples from archaeological excavations), but shorthand was also used on papyri and parchment.

Though his primary training was in computer science, Baumann has been working with the Duke Collaboratory for Classics Computing since 2013 and on papyri since 2007. The Duke Collaboratory runs papyri.info, an open access online resource that gathers information about ancient Greek papyrus manuscripts and their contents. Duke is one of the foremost institutions in the world for working on ancient manuscripts—it not only has a remarkable collection, it is the home of numerous prominent manuscript experts. So, Baumann was in a position to think more about these ancient codes. […]

Unlike the forms of shorthand that dominated 20th-century Europe and America, however, ancient Greek and Latin shorthand was not always standardized and was uncompromisingly difficult to learn. A second-century contract from Egypt tells us that it took two years for a literate enslaved child to become proficient in Greek shorthand. The curriculum began with the student memorizing a basic set of signs vowels, syllables, word endings, and phrases before progressing to a more complicated system of compound signs. In this second step a single sign was modified by a dot or dash to augment its meaning. With only a few exceptions, these symbols were in no way pictographic, so you couldn’t guess what they meant just by looking at them. This made it useful for military communications and espionage: Both Julius Caesar and the Jewish freedom fighters who led the mid-second-century Bar Kochba revolt used it to send messages.

Those who learned ancient shorthand were aided by something called the “commentary,” a set of sentences that served as mnemonic devices for the student. A copy of this commentary was published by Sofía Torallas Tovar and Klass Worp in the important volume To the Origins of Greek Stenography. It has helped scholars decipher more complex signs and understand how it was learned. “At the conclusion of this process,” said Baumann, “the student would have learned a system of over 800 signs alongside what is essentially a gigantic lookup table of over 4,000 words and phrases.”

In my own work I have argued that, to the best of our knowledge, only those who went through this arduous process—that is enslaved and formerly enslaved workers—could actually read it. To the untrained eye it resembles squiggles or chicken scratch. Given the transformational nature of translating the spoken word first into symbols and subsequently into ancient Greek (or Latin) literacy seems important. (Full disclosure: I have an article coming out early next year in the Journal of Theological Studies in which I discuss the ramifications of this for thinking about the composition of early Christian texts and the ways in which enslaved people might have resisted structures of power). […]

The problem, Baumann told me, is that scholars haven’t been that interested in shorthand and, thus, haven’t been fully or accurately noting its presence. When an editor just notes that there were “shorthand marks” in a particular ancient text this doesn’t tell us even how many there were much less what they were. Sometimes editors of manuscripts describe it not as shorthand, but as an “unidentified script.” If an entire text was written in shorthand, it might have been misidentified at some point in its study. And, in some cases, editors might not have registered it at all. The lack of editorial consistency makes it difficult to get a sense of how much untranslated shorthand there even is. Short of going back and reviewing the hundreds of thousands of papyrological fragments sat in university libraries, museums, and private collections we can’t really be sure what is out there.

This is one space where digitization could be helpful if additional resources were available to scholars. We lack the adequate tools for notation: we don’t have a font for shorthand script so even if you wanted to transcribe ancient shorthand and print it in a book or put it on a website like papyri.info you wouldn’t be able to. In 1992, Giovanna Menci started a database of papyri that included shorthand with the goal of eventually establishing a corpus of shorthand texts. But the project was abandoned due to technological challenges and a shortage of funding.

What this means, then, is that there are codes inscribed in ancient manuscripts that we could read but aren’t. This material might give us insight into the hidden spaces of ancient writing culture and access to scientific thinking, bureaucratic processes, and literary interpretation. It’s especially promising for helping us understand the obscured lives of enslaved people and decipher the workings of the Roman military. But, currently, only a select group of scholars and librarians are even thinking about it. The first steps, Baumann said, are to produce transcriptions and readings of the papyri we have and expand our understanding of how it worked. This is still an obscure sub-sub-discipline and just gathering what we know so far would allow a larger number of people to participate in the project. If we want to understand stenography, Baumann told me, “We must first reanimate the dead understanding of ancient Greek shorthand.” If we do that then “knowledge that was lost for over a millennium can be reclaimed.”

There’s more at the link; I couldn’t help but note the oddity of “codes […] that we could read but aren’t” — I would have queried “but don’t” or “but aren’t able to” had I been copyediting it.

Comments

  1. David Eddyshaw says

    I knew about Tiro, Cicero’s freedman, using shorthand, but kinda assumed it was his own invention. Seems I was not alone in this misconception:

    https://en.wikipedia.org/wiki/Tironian_notes

  2. Well, he probably let people think it was. Tricksy, these underclass folks…

  3. “but dictation was also used to compose letters, philosophy, and narrative”
    Sigh. I need a secretary and I need philosophy.

  4. David Eddyshaw says

    It’s pretty impressive that Cicero, Mr Style ipsissimus, could actually dictate Ciceronian periods; but then, I suppose that’s what being a great orator does for you.

  5. That Cicero, Mr Style ipsissimus, could actually dictate Ciceronian periods (but then, one might suppose that that is what being a great orator does for one), as recorded in the obiter dicta of a frequent correspondent who rejoices in the name (among others?) of “David Eddyshaw”, is pretty impressive. (Someone write that down, please.)

    I couldn’t help but note the oddity of “codes […] that we could read but aren’t” — I would have queried “but don’t” or “but aren’t able to” had I been copyediting it.

    Engaging oddity indeed. A fuller extraction from our text:

    1. “There are codes that we could read but aren’t.”

    Is there some construal that makes this grammatically acceptable (passable by grammar checkers such as MS Word’s, furnishing one explanation but not an excuse)? Yes:

    2. “There are codes that we could read but that we are not [identical with].”
    (So trivially true that it could not be what was meant.)

    Word for Windows 365, as standardly configured, accepts 1 and also 2. But then, it is so unciceronian that it accepts as grammatical the quasi-sentences *”These are codes that we could read but isn’t.” and *”Which codes we were eat but.”

    Hat’s query concerning “but don’t” is of course apt; “but aren’t able to” is trickier:

    3. “There are codes that we could read but aren’t able to.”

    Grammatical, but is it likely to be what was meant? I doubt it. Among the candidate meanings:

    3a. “There are codes that we could [formerly] read but can’t [any more].”

    3b. “There are codes that we could read [if circumstances differed] but can’t [given the way things are].”

    Another “correction” to consider:

    4. “There are codes that we could read but aren’t reading.”

    That might be what the author had in mind: simply dropping “reading” on the assumption that it was not grammatically needed. Quite plausible. Anyway, I like “but don’t” best of all.

  6. One more variant:
    *could be read

  7. Dravsi, that’s a bit cryptic. Perhaps you meant this:

    “There are codes that could be read but aren’t.”

    Or something else, perhaps?

  8. and enoch read [the codes] with god and wasn’t.

  9. this is making me wonder about the relationship of dictation to :

    1) the production of the homeric texts. the origins are almost certainly in semi-improvised, episodic performance, but there’s a certain haze between that and the existence of the oldest versions of the written text. there’s a certain assumption that the haze is incarnated as an ionic elias lönnrot, but knowing that there was a widely-used dictation shorthand around on the one hand makes that seem a more literal possibility, and on the other makes more plausible the idea that the extant text is a more purely literary creation assembled with no direct connection to the oral tradition, but by mining an array of manuscripts of many ages and provenances produced through transcription of different performers’ tellings of the epic.

    2) the production of the qu’ran. gabriel’s command “recite” means something different in a world where it can mean “transmit this text once, so that it can be accurately recorded by a professional transcriber” as well as “speak this text to confirm that you have memorized it”.

  10. knowing that there was a widely-used dictation shorthand around

    We don’t know that, though. I don’t know when the earliest attested papyrus shorthand dates from, but it must be several centuries after the Homeric poems were first written down. It sounds like a Hellenistic thing.

    It would be very interesting to read “the commentary” — I wonder if it’s online somewhere.

    “Codes that we could read but aren’t” sounds standard to me.

  11. January First-of-May says

    Another “correction” to consider:

    4. “There are codes that we could read but aren’t reading.”

    That might be what the author had in mind: simply dropping “reading” on the assumption that it was not grammatically needed. Quite plausible. Anyway, I like “but don’t” best of all.

    I agree fully. This does seem to match the context, and it does appear that don’t is the grammatically correct replacement in this meaning.

     
    An unrelated point…

    Often this would have happened on erasable wax tablets (we have many examples from archaeological excavations)

    This is strangely phrased; are there actually any archaeologically excavated Greek wax tablets sufficiently well preserved to tell which script they were written in?

    (It’s not a priori impossible, because if we drop “Greek” the answer is in fact positive, though I’m only aware of one such example. In principle it could also be possible to read wax tablet text from impressions left on the non-wax part, but I’d imagine it would be almost impossible for shorthand in particular.
    But AFAIK the actual situation is that archaeologists find 1. many illegible tablets and 2. many texts on other media mentioning wax tablets as something people write on.)

    [EDIT: apparently yes, there are many examples. TIL. Wouldn’t have expected.]

  12. David Eddyshaw says

    the obiter dicta of … “David Eddyshaw”

    All my dicta are obiter. I lack your human concept of “relevance.”

  13. For me,

    “There are codes that we could read but aren’t.” — ungrammatical

    “There are codes that we aren’t reading but could.” — grammatical

    Interesting. I would guess that some readers find the latter unacceptable but not as bad as the former?

  14. David Eddyshaw says

    Grammatical for me, though with a bit of delay in the parsing. I think I myself would actually say “… but aren’t doing”, but I find the el|iptical version perfectly acceptable (though clunky.)

    IIRC there is a known UK/US difference over dropping forms of “pro-verb” do, but I’ve forgotten the details.

  15. MM:

    “There are codes that we could read but aren’t.” — ungrammatical

    Yes, interesting. We note that TR finds it “standard”. Does TR mean that people utter such things often? I would accept that limited claim, but not much beyond it. And David Eddyshaw finds it “perfectly acceptable”.

    Sensibus strictis: I don’t say that it’s “ungrammatical” (a contestable term), just that if it’s parsed in a way that makes it likely to be judged “grammatical” then it can’t mean what it was intended to mean (see above).

    A further candidate amendment:

    5. ?“There are codes that we could read but that aren’t.”

    If we take each that in 5 to be a relative pronoun (quaintly, and pace Pullum and co), it does seem to lack one puzzling difficulty found in the original:

    1. “There are codes that we could read but aren’t.”

    In 5 the first that is the object of “could read” and the second is the subject of “aren’t [read]”; in 1 the single that must fill both roles. Makes most of us queasy.

    To explore further I would substitute a regular verb for read, and try for a convincing parallel that does without the modal ambiguities of could, issues with that, forms of to be that are not common to we and to they [the codes], and so on. Strip it down to essentials. Trying to do so reveals the slippery outlandishness of 1:

    6. *?”Some codes I can understand but aren’t.”

    7. *?”This code I understand but isn’t.”

    8. *?”There exists some code that I understand but it isn’t.”

    The exercise seems futile; and that futility suggests (unfutilely) that the crux is indeed located in the that, with further obscurity lent by sundry uncertainties or bivalencies: a single read with two parsings; a single that with two parsings; even a single aren’t with two parsings.

    This expansion of 5 seems to do all that could be asked, by removing bivalencies:

    9. “There are codes that we could read but that aren’t read.”

  16. Jen in Edinburgh says

    I don’t have a problem with ‘codes we could read but aren’t’, although I might have written ‘could be reading’

    For 6, I’d have to substitute ‘don’t’ just because of the awkwardness of ‘I aren’t’, but I’m not sure that’s the same issue?

  17. David Eddyshaw says

    Looking it up in CGEL, I see that my confused recollection that there are UK/US differences in the dropping of “pro-verb” do is correct, but that it doesn’t actually bear on this issue.

    CGEL treats the ellipsis of the actual -ing form of the verb in the progressive on p1521, under Ch 17, 7.1 “Stranding of auxiliary verbs”, and describes constructions like this as “of somewhat marginal status”; the examples they give are

    ? “KIm won’t enter the competition, but Pat is.”
    ? “They may all move south, and in fact some of them already are.”

    They say that the adjunct “already” in the second example improves it slightly, but that “fuller marking of the progressive (e.g. is/are doing so) would generally be preferred.”

    In other words, this is yet another case where there is no neat line between grammatical and ungrammatical. Take that, Noam!

  18. Lars Mathiesen (he/him/his) says

    I’m not a native speaker, but I did spot this in the OP and my initial reaction was “What? Oh, it used to say be reading and they changed it without noticing they needed to change the conjunct.” So: the meaning is clear, but it gives pause and makes me think about grammar which is not what the writer wants.

  19. And David Eddyshaw finds it “perfectly acceptable”.

    No, no, what he found “perfectly acceptable” was the second option, “There are codes that we aren’t reading but could.” I find I have the same reaction to it as DE. Odd!

  20. David Eddyshaw says

    No, I did actually mean that I found the “aren’t” version grammatical, although it took me a few microseconds longer than usual to supply the ellipsis as intended. Hence, grammatical but clunky. It seems to be parallel to CGEL’s debateably-grammatical examples, so it’s not surprising that even native speakers’ intuitions about it vary.

  21. Oh! It was I who completely misread!

  22. Both “There are codes that we could read but aren’t.” and “There are codes that we aren’t reading but could.” are grammatical to me. The first to me is more neutral (are we not reading them because we don’t have the knowledge, or because there are institutional/structural reasons preventing us from exercising said knowledge), whereas the 2nd one I would invariable read as meaning that we have the intellectual/physical capabilities to read the codes but cannot due to some reason external to the codes themselves and we must be exhorted to overcome those external barriers.

    Randomly swapping pairs of nouns and verbs continues to retain the grammaticality for me: there are rockets that we could fly but aren’t, there are books that we aren’t making but could, there are widgets that we could code but aren’t, there are songs of sixpence that we aren’t singing but could, etc.

  23. In other words, this is yet another case where there is no neat line between grammatical and ungrammatical. Take that, Noam!

    Did Old Nick genuwinely hold that there was a neat line ? That would be the stupidest opinion I’ve encountered this week. No need to go on further into bindings and gummint, or whatever they are.

    It does give me pause that, as I just realized, I now measure stupid opinion frequency per week rather than per month, since the figures were getting too large. Despite reading less and less news ! I must be getting old. It’s not that I think *my* opinions are non-stupid, I have after all very few and I fine-tune them as needed.

  24. @Noetica, yes, I meant that technically “but aren’t” could be a part of this construction.


    The comparativist asterisk means “unattested”, and I think when I first encountered the generativist asterisk (which as far as I understand means: “at the moment I really don’t like this construction” (I’m not sure if “grammaticality” is a thing)) I just expanded “unattested” to “hypothetical” (as a concept that can meaningfully cover both the generativist and comparativist usages).

  25. Not quite. The diachronic asterisk is “unattested, but likely existed”. The synchronic asterisk means “wrong (and therefore unattested)”.

  26. The diachronic asterisk is “unattested, but likely existed”.
    Exactly. That’s why one finds double asterisks for forms that are unattested and didn’t likely exist, e.g. for rejected reconstructions.

  27. diachronic – synchronic
    I considered using these two words, but the generativist use is different from “unattested in the corpus of modern texts”…

    but likely existed
    I think, “if [some sound law were true] then the Italic reflex of *[an IE form] would be *[…] rather than *[…]” is fine.
    And even synchronously:
    “…. then the modern reflex would be *[…] rather than […]” is also fine.

    P.S. Yes, the double asterisk. But is it used systematically? And was it used origianally?

  28. I can be wrong, but I thought that the contrast was (at least at some point) between real/observed forms and theoretical constructs, and that * / ** is a secondary distinction.

    In this respect, *I loves Masha is indeed somewhat more theoretical than I love Masha for me, even though I don’t know any Masha closely.

  29. January First-of-May says

    The comparativist asterisk means “unattested”, and I think when I first encountered the generativist asterisk (which as far as I understand means: “at the moment I really don’t like this construction” (I’m not sure if “grammaticality” is a thing)) I just expanded “unattested” to “hypothetical” (as a concept that can meaningfully cover both the generativist and comparativist usages).

    I agree, the concepts are very similar. I don’t recall which I encountered first, but whatever it was, I suspect I made a similar conceptual leap to the other.
    Of course comparativist/diachronic notation has no equivalent for the generativist/synchronic question mark (for sentences that fall in the gap between clearly grammatical and clearly ungrammatical).

    As far as I’m concerned, the double asterisk is for counterfactual reconstructions, usually in a context to the effect of “then the reflex would be **[…]” (here’s an example from this Monday), though I think I’ve also seen it used for merely particularly implausible reconstructions (and might well have used it that way myself, but it’s not very easily searchable across multiple comment threads).

    I vaguely recall a LH thread from last year (or so) that went into some detail on the history of the double asterisk, and on other marks used for counterfactual reconstructions (I think one was some kind of superscript X thing). Unfortunately, as I just said, it’s not exactly easy to search [and there’s just two Google hits on LH for the actual phrase “double asterisk”, one of which isn’t even about reconstructions… I tried “double star” but that’s swamped by astronomy*].
    IIRC the consensus was double asterisk is a fairly recent introduction, but it was one that quickly caught on for its convenience.

    EDIT: I wonder if the counterfactual double asterisk developed through contexts like this textbook quotation [“the initial resonant clusters like *wr- (not **ur-) or *ml- (not **m̥l-)”, and many other examples within the paragraph], where the comparativist “this is a reconstruction” asterisk is apparently combined with the generativist “this is not the actually used form” asterisk.
    [Found this while searching for an implausible-reconstruction example I thought I’ve seen, which actually only had one asterisk and wasn’t that implausible anyway.]

     
    *) meanwhile, the only hit for “two asterisks” is about footnotes, and “two stars” is again all astronomy

  30. ** sometimes also means a double hypothetical, e.g. Nostratic reconstructions based on PIE, PU, etc. reconstructions.

    By “wrong” I meant the generative sense of “judged unacceptable”. In principle that is not the same as “unattested in the corpus”, because competence/performance.

  31. I find, “There are codes that we could read but aren’t,” clearly comprehensible but glaringly ungrammatical. Following, as it did, the weird-sounding “sat” in, “Short of going back and reviewing the hundreds of thousands of papyrological fragments sat in university libraries, museums, and private collections we can’t really be sure what is out there,” I found myself wondering about whether there were systematic gaps in the author’s English language competence.

    As to the possibility that the Homeric epics might have been recorded in shorthand, before they were standardized in written Attic Greek, it seems unlikely. The shorthand methods appear to date from much later, and the glyph forms used in the shorthand seem to suggest that they were developed by people who were already familiar with a Phoenician-derived script. The content of the epics also points in this direction. There is no indication that Homer (or whoever set the works in something resembling their final forms) had any awareness of writing whatsoever; writing, whether in a Phoenician-derived script or a (period-accurate) Linear script is simply never mentioned. Compare this with the heroes’ use of chariots; Homer at least knew that the nobles of the time rode chariots, even if he had no real conception of how they might actually have been used in a pitched battles.

  32. Yes, there is a difference here.

    It seems many ascribe the asterisk to Schleicher, but that is dubious. Cf. also E. F. K. Koerner, L’asteriso nella linguistica storica :

    The preceding note constitutes but a partial history of the origin and development of (the use of) the asterisk in linguistics, whose locus classicus is no doubt August Schleicher’s Compendium of 1861 although Schleicher was by no means the inventor of the either the symbol or the concept. There is no clear indication that H.C. von der Gabelentz and Julius Loebe introduced the asterisk into linguistic nomenclature when they made regular use of ‘starred forms’ in their Glossarium der Gothischen Sprache of 1843. It appears much more likely that they found the procedure to mark unattested forms with a preposed asterisk already in use in other linguistic writings of the time, most likely in lexicographical work since reconstructions of Indo-European proto-forms à la Schleicher were not undertaken before 1850.

  33. the timing of the appearance of the shorthand seems to me to be an argument against an assembly-from-transcribed-performance-corpus process for the homeric epics.

    but i’m not sure that the lack of references to writing tells us much. aside from some specific subgenres*, i’d be hard put to name more than a few yiddish folk songs that mention writing (only one comes to mind immediately: “shtey ikh mir afn ganikl”**), and that’s a body of work collected in the 19th century and later in communities where literacy was widespread (though far from universal, despite persistent pseudo-ethnographic fantasies). literary forms that are primarily written are often obsessed with writing (from arabic letters as reference points for beautifully-shaped facial features in andalusi love poetry on down), but i don’t know that i’d expect references to writing to show up in primarily oral forms just because it’s around and familiar as a technology.

    .
    * maybe only two, even: emigration/conscription letter songs; songs for/about boy-children that define success as talmudic learning.

    ** which might be a version of a conscription letter song, given the figure of the bird as messenger of the death of the beloved (here bringing a letter, as opposed to speaking as it does in conscription letter songs like “af di grine felder un velder”/”dos fertsente yor” a/k/a “chorna rillia izorana” in its ukrainian version)

  34. Does TR mean that people utter such things often?

    Yes, but also that they occur fairly often in more-or-less informal writing styles like that of the quoted article. I’m surprised so many Hatters find the sentence objectionable, frankly — it doesn’t give me pause at all.

    In 5 the first that is the object of “could read” and the second is the subject of “aren’t [read]”; in 1 the single that must fill both roles. Makes most of us queasy.

    But the elision is “aren’t reading”, not “aren’t read”. Noetica’s 5 (There are codes that we could read but that aren’t”) is completely ungrammatical to me precisely because it rules out the “reading” reading.

  35. Brett, ‘the weird-sounding “sat”’ is not a “gap in the author’s English language competence.”; it is increasingly common colloquial British. See https://www.grammarphobia.com/blog/2017/12/sat-stood.html

  36. What a world where every conversation that happened 200 and more years ago is recorded and accessible would look like?

  37. TR:

    Noetica’s 5 … is completely ungrammatical to me precisely because it rules out the “reading” reading.

    Voyons:

    5. ?”There are codes that we could read but that aren’t.”

    Yes, 5 rules out an interpretation involving “[we] aren’t reading”; but does that make it ungrammatical? Hardly. I expanded 5 as follows:

    9. “There are codes that we could read but that aren’t read.”

    By your reasoning that too would be ungrammatical (“precisely because it rules out the ‘reading’ reading.”). But does anyone here think it is? Segue to an obiter dictum of David Eddyshaw:

    “In other words, this is yet another case where there is no neat line between grammatical and ungrammatical. Take that, Noam!”

    Quite. As I myself had said (sensibus strictis):

    “I don’t say that it’s ‘ungrammatical’ (a contestable term), just that if it’s parsed in a way that makes it likely to be judged ‘grammatical’ then it can’t mean what it was intended to mean.”

    Relevant here (sorry David, this part is for the rest of us) is the swampy no-analyst’s land between the “ungrammatical” and usages deemed improper by so-called prescriptivists – and sometimes militantly advocated by so-called antiprescriptivists (many of whom I call cryptoprescriptivists) on account of their supposed “grammaticality”. And on account of their being in common use. See this ancient comment (the asterisks), and follow-ups in that thread.

    What a world where every conversation that happened 200 and more years ago is recorded and accessible would look like?

    (I love it!)

  38. There is one reference to writing in Homer, in Iliad 6 where Bellerophon is sent to the king of Lycia with a tablet bearing σήματα λυγρὰ … θυμοφθόρα “baleful, live-destroying signs”, i.e. a message to the king to kill him. I don’t know what scholarship thinks about the date of composition of this episode.

    I think if I came across There are codes that we could read but that aren’t in an English text I’d do the same double-take that Hat did with the original version. The filled-out version There are codes that we could read but that aren’t read is of course fine. What I meant was that the possibility of reading “reading” makes 1 unproblematic for me, and that if it were possible to read that in 5 then 5 would be fine too.

  39. David Eddyshaw says

    mollymooly is (naturally) quite right: “sat” in that sense is perfectly normal colloquial UK English.

    The linked grammarphobia article is fairly seriously misguided: for one thing, “sat” is being used here to mean “seated”, rather than “sitting”; this has nothing whatsoever to do with any imaginary “decline of continuous tenses.”* Needless to say, it’s not an “error”, as the ill-informed Catherine Soanes (Soames, more likely, as in the second invocation) is cited as saying. Tell it to the French … and indeed to the multitude of speakers of other languages that do this sort of thing regularly with body-position verbs.**

    * I see that the grammarphobia blogger has in fact correctly worked this out, unlike Soan/mes.

    ** Please note that I am heroically resisting the temptation to explain how they work in Kusaal (or even Swahili.)

  40. David Marjanović says

    There is one reference to writing in Homer

    …but it reads like (this particular) Homer was thinking of magic signs and did not actually have an idea what writing had been like.

  41. The signs are life-destroying in that the king reads them and tries to kill Bellerophon; they’re not magical in themselves.

  42. It’s Catherine Soanes with an N (yes, Grammarphobia got it right once and wrong three times), and shame on Oxford Dictionaries for dumping the blog in the trash — that post is gone, it wasn’t even captured by the Internet Archive.

    The same post by Soanes was previously cited here by AJP Crown in support of his own belief that “I was stood” was exactly the same as “I was standing”. However, he was interpreting it from the outside: “I don’t use it myself, but I’ve spoken to enough people who do, mostly northerners, to know that there’s nothing implied that wasn’t in the -ing form.” And he seems to have been outvoted in that thread by Lars (the original one), mollymooly, and AntC, who all felt there was a difference, though the first two are also outsiders who don’t say it themselves, and I’m not sure about AntC.

  43. January First-of-May says

    The signs are life-destroying in that the king reads them and tries to kill Bellerophon; they’re not magical in themselves.

    Perhaps Homer (or whoever helped him compose that particular story) had heard of writing but thought it was a Lycian thing? Or, to be more precise, had encountered reports of some eastern barbarians using strange signs to write down messages.

    Specifically, AFAICT the Luwian hieroglyphs were in active renaissance in Homer’s time, though admittedly mostly in places well east of Lycia. (Shortly afterwards they were replaced by Phoenician and its derived scripts even there.)

    [EDIT: of course it’s also possible that the story is just very old, itself, and Homer just repeated it without understanding how that worked. Are there any older examples of the same plotline? I know it from Hamlet, mostly, but of course that could have been derived from the Iliad itself…]
    [EDIT 2: apparently there is in fact a Sumerian example, so yeah, it’s an old plot.]

    that post is gone, it wasn’t even captured by the Internet Archive

    It was in fact successfully captured on an earlier address. The comments are gone, though.

  44. “It was in fact successfully captured on an earlier address. ”

    Your google-fu is stronger than mine today:(

  45. Wow, thanks, JFoM! Impressive! I wondered if the context was as bad as the bits quoted at Grammarphobia — turns out, it’s *worse*:

    Aarrgghh!!! … (being a sensitive soul when it comes to incorrect grammar) it gets to me just as much as hearing four-letter words on daytime radio. … why, oh why do many people say ‘I was sat’ or ‘we’re stood’? … I hope this blog has helped those of you who (perish the thought!) use this construction to see the error of your ways.

    Wow, that’s unprofessional. Camping it up in the hope that the camp covers up the class/regional prejudice (perish the thought!).

  46. David Eddyshaw says

    “Four-letter words (indeed, any obscenities) aren’t standard English” says la Soanes …

    How did this person ever gain employment as a “lexicographer”?
    (The rubbish about tenses also displays fundamental grammatical ignorance.)

  47.  I’m not sure about AntC.

    AntC isn’t so sure himself, now that he’s not visited Yorkshire since before there was sensible government.

  48. Soanes is identified as an “ex-lexicographer” at the top of the blog, so there’s that.

  49. In We were sat in a pub having a drink, that “sat” could be replaced with “seated” (or could it? I’d want to ask whoever said it whether they’re the same). But what about the hundreds of thousands of papyrological fragments sat in university libraries? I don’t think a papyrus can be seated. Is it maybe an implied causative, i.e. caused/made/forced/had no choice but to sit/stand, and to remain in that position? (As an American, I have no real clue, just throwing out guesses.)

    Notes on “sat”/“stood” were added in Burchfield’s revision of Fowler’s Modern English Usage in 1996. Burchfield thought they were acting as present participles:

    sat, used as a present participle (= standard English sitting), has been rediscovered by dialect scholars, who say that it is widely used in parts of the north and west of England. The following examples seem to confirm that its currency on the fringes of standard English is increasing: I can’t help thinking of that Tim sat there juddering his leg up and down—K. Amis, 1988; Now, I’m sat in a nice car, my husband at my side—A. Duff, 1990 (NZ); now, as a result of a conference débacle, you are sat on the back benches with nobody wanting to sit next to you—B. Elton, 1991. The use is exemplified in the OED s.v. sit v. 18b, with examples from OE down to 1864. It is firmly marked ‘Now dial.’ In other words it is an example of a use that was once standard but has gradually become regionally restricted over the centuries. Cf. STOOD.

    stood, used as a present participle (= standard English standing) presumably has a dialectal distribution in BrE similar to SAT, but if it has its history seems to have gone unrecorded in the OED. Its existence in modern regional use is not in question, but its precise distribution has not been established. J. Cheshire et al. (1989) conclude that it and the similar use of sat ‘are now becoming characteristic of a general nonstandard or semistandard variety of English’. Examples: She was stood in front of the mantelpiece trying to think of the name for the clock—A. Bennett, 1981 (Yorkshire); ‘But that’s not the half of it.’ Uncle Simon sat forward. ‘Do you know what he did when he was stood there face to face with the priest, the man who positively identified him?—G. Patterson, 1988 (NIr.); And she’d pay the driver, and she’d be stood there, on the soiled concrete footpath—A. Duff, 1990 (NZ, Maori speaker); My husband was stood on the opposite side of the pitsCycling Weekly, 1993.

    Jenny Cheshire in various publications also calls these “present participle sat” and “present participle stood”, but maybe that’s just a shorthand label? I wonder if there’s disagreement in the literature on whether that’s what they are.

  50. David Eddyshaw says

    what about the hundreds of thousands of papyrological fragments sat in university libraries?

    Good point. More epicycles needed …
    I think “sat” here is functioning like a predicative-only adjective, like “asleep”; “situated” seems to cover the relevant meanings.

    It occurs to me that this “sat” differs from “seated” in another way: it requires a complement. You can say “I’m sat at the table” or even (I think) “I’m sat quietly” but not just “I’m sat.” Same for “stood.” (It differs from “sitting” in this respect, too.) It’s clearly not a participle; you can’t (for example) say *”a sat duck.”

    Soanes is correct in saying that it’s not a passive: in “I’ve been sat by those people”, “by” must mean “next to”, and not “by the agency of.”

    However, I don’t it’s a stretch to connect this with constructions like French je suis assis or to the Swahili nimekaa “I am sitting”, where the -me- marks so-called “perfect tense”: nimesoma “I have read” (and the verb is reconstructable to Proto-Volta-Congo: Proto-Bantu *kada, Proto-Oti-Volta *kaʎ-, cf Moba kal.)

  51. The 2015 revision of Fowler (by Jeremy Butterfield) has further developments, and explicit warnings of “best avoided”, indicating it’s now common enough to cause anxiety:

    sat, combined with the auxiliary be in place of a progressive tense (i.e. she is/was sat, instead of she is/was sitting), has become a topic of interest because of its increasing use: it raises the question of what is ‘standard’. The OED entry for sit (not fully updated) shows that this use stretches back to Old English (earliest citation, Anglo-Saxon Chronicle for 922), and the editors of the entry (1911) labelled it ‘dialectal’; the latest citation they showed (1864) was from a book of Lancashire rhymes and clearly dialectal. So much for the history. For many people from a large part of the North and part of the West of England, it continues to be the normal formulation, and they might well be offended to hear it described as ‘dialect’. What has changed is that, as the OEC data suggests, its use has spread well beyond its traditional regional confines; though still a predominantly British construction, it is also found in other varieties, including AmE, Australian, NZ, and Irish English. Unsurprisingly, it appears above all in texts classified by the OEC as not being formal or technical, especially in blogs, news, and sports writing. Many of the British newspaper examples are from Northern papers, often directly quoting what people said. However, even a paper as starchy as the Telegraph occasionally slips it in, as in the last example below. For the moment it is likely to raise eyebrows in some quarters, so is best avoided in any kind of writing where formal standard English would be expected. Examples: [same as previous edition, plus:] It means that if people are sat here waiting a long time they will see new things each time they look at itThis is Wiltshire, 2004; Both Mr Swinney and Mr Russell were sat near the First Minister as he misled parliament by claiming that the £546 million college budget increased this year— Daily Tel., 2012. For a parallel phenomenon, cf. STOOD.

    stood, used, like SAT, with the auxiliary be in place of a progressive tense (i.e. she is/ was stood, instead of she is/was standing), presumably has a regional distribution in BrE similar to sat … [continues as in previous edition, then adds:] The types of register in which it appears are the same as for sat, and, just like sat, it appears often in journalism from the North of England reporting what people said. It seems to be very much a spoken form, and is best avoided in any kind of standard formal prose. Examples: [same as previous edition]

    (“Progressive tense” [sic].)

  52. David Eddyshaw says

    Actually, come to think of it, “I’ve been sat by those people” does permit a passive interpretation, even though the unnamed individuals who have caused me to sit by those people cannot be those people themselves.
    I would usually interpret it in isolation as meaning “I’ve been sitting by those people”, though.

  53. You can’t exclude obscenities as in Kama Sutra positions, but you can, I think, exclude unprintable words.

  54. David Eddyshaw says

    Well, it’s true that if you can’t print it in the dictionary, you can only allude to it obliquely …

  55. Here I imagined a dictionary where four-letter words are described (and defined) in the style of Arabian Nights.

  56. David Eddyshaw says

    Well, there are always pictures

  57. David Eddyshaw says

    Butterfield’s reference to the Anglo-Saxon Chronicle is interesting. However, the use of “be” rather than “have” to form the perfect of intransitive verbs is regular in Old English, so this would simply be a straightforward case of exactly the same sort of thing as the Swahili nimekaa “I’ve sat down” ↔ “I am sitting” (and abundant cross-linguistic parallels.) If the construction has been in continuous use since Old English (which is not altogether clear from what B actually says) then “I’m sat/I’m stood”, far from being Horrid Neologisms, would be Venerable Survivals, worthy of assiduous imitation by all Soaneseses.

  58. “I’ve been sat by those people” does permit a passive interpretation

    But isn’t that a different construction? All of the examples so far have been “to be sat”, not “to have been sat”. Also, in “to have been sat”, it would have to be the same as “seated”, wouldn’t it? i.e., you can’t have *papyrus fragments that have been sat in libraries.

    The argument that this isn’t a passive because there’s no such thing as “to be sat/stood by (some agent)” sounds pretty decisive to me. I wonder if that criterion is discussed in the literature. And the parallel to “asleep” is strong, too: you enter a bodily state/position and remain in it for some time.

    Is there any difference between asleep and sleeping, other than where you can put them syntactically? That is, you can’t have *an asleep child and you can’t *fall sleeping (and maybe a few other cases), but otherwise they’re interchangeable. So if that’s the analogy, it would seem to imply that “to be sat” is the same as “to be sitting” … on the other hand, you could argue that the -ing form suggests a process or activity while the non-ing form suggests a state, if that’s really a difference.

  59. The OED has also now weighed in on “to be sit/stood”. They call them “passive” and “equivalent to the progressive.” From sit, revised December 2020:

    27c. transitive (in passive). to be sat: to be sitting, to be in a seated position. Cf. to sit down 2b at Phrasal verbs 1. Now British colloquial.

    In the 19th and 20th centuries, regional and nonstandard, but increasingly common since the late 20th century. …

    [last quotations:]
    1981 J. Sullivan Only Fools & Horses (1999) I. 1st Ser. Episode 4. 37 (stage direct.) Pauline is sat at a table.
    2017 @carolinefoste5 23 Oct. in twitter.com (accessed 9 Nov. 2017) I should be doing work but instead I’m sat on my bed watching Strictly.

    to sit down 2b. transitive (in passive). to be sat down: to be sitting down, to be in a seated position. Cf. main sense 27c. Now British colloquial.

    Apparently rare from the 19th century until the late 20th century. …

    [last quotations:]
    1986 alt.folklore.computers 16 Dec. (Usenet newsgroup, accessed 24 Nov. 2017) James, our manager, was sat down, head in hands, hands between knees.
    2012 K. Hudson Tony Hogan bought me Ice-cream Float 177 You could see an inch of grey sock and three inches of greyer skin when l was sat down.

    and stand, revised June 2022:

    5g. transitive (in passive). Chiefly British colloquial. to be stood: (equivalent to the progressive, usually in senses 5a(a), 5d) to be continuously in a standing position in a place or while performing an action; to be standing.

    In the 19th and 20th centuries, regional and nonstandard, but increasingly common since the late 20th century. …

    [last quotation:]
    2003 Independent on Sunday 30 Mar. (Review Suppl.) 14/1 I was stood waiting to get my pay-cheque cashed.

    I’m still wondering if there’s disagreement in the literature over the “passive” label.

  60. David Eddyshaw says

    But isn’t that a different construction?

    Yes, potentially: it’s ambiguous between a passive reading (which is a different construction) and a perfect active (which is the same construction but with a different tense/aspect.)
    Cf

    “I’ve been stood in the corner for being a prescriptivist.” (passive)
    “I’ve been stood here for an hour now.” (perfect active)

    The first type could in fact have an agent supplied: “I’ve been stood in the corner for being a prescriptivist by that gang of rabid descriptivists”; my criterion didn’t work as I proposed because I missed the ambiguity in my example sentence. But Soanes in nevertheless correct in saying that the construction she deplores is not passive:

    “I’m sat by those nasty prescriptivists.”

    can’t be interpreted with “by” meaning that the prescriptivists caused me to sit, but I think this is because “sat” needs a complement, and the by-phrase is the only candidate available. I think you could just about read

    “I’m sat here by those nasty prescriptivists.”

    in the sense that the nasty prescriptivists habitually cause me to sit in this place; this restores the potential ambiguity.

    “I’m sat here by those nasty prescriptivists every second Tuesday.”

    This remains ambiguous, and can also be interpreted in the habitual sense “I sit here by those nasty prescriptivists every second Tuesday.” In other words, the “sat” construction does not necessarily have continuous meaning; and this does not mean quite the same as “I’m sitting here by …” The idea that “sat” is a substandard substitute for “sitting” in these constructions is quite untenable.

    EDIT on seeing your last post: the OED is surely wrong in calling this “passive”: it actually contrasts with the passive uses.

    Does the OED record the construction as having been in use all along since Old English up until the Age of Misinformed Prescriptivism, or did it disappear and resurface in dialect more recently?

  61. the use of “be” rather than “have” to form the perfect of intransitive verbs is regular in Old English

    Yes, the revised OED brings up that point—but they *don’t* take it back continuously to Old English. I snipped out their comment on the perfect from the quotations above because they were so long; here’s what was cut:

    sit 27c. to be sat
    Some Middle English and early modern English examples may show either the perfect tense formed with be rather than have or the passive of sense 1b (as in we being sat ‘we, having sat down’ or ‘we, being caused to sit’).
    c1300 Childhood Jesus (Laud) l. 1720 in C. Horstmann Altengl. Legenden (1875) 1st Ser. 57 To þe bord huy beoþ i sete.
    c1380 Sir Ferumbras (1879) l. 48 Al on murȝþe was he y-sete wiþ a fair baronye.
    c1480 (▸a1400) St. Matthew 9 in W. M. Metcalfe Legends Saints Sc. Dial. (1896) I. 190 In þe tolbuth set lewy, þat as a tollare þare wes sate.
    1609 W. Shakespeare Louers Complaint in Sonnets sig. K2v Hee againe desires her, being satte.

    to sit down 2b. to be sat down
    Some early modern English examples may show either the perfect tense formed with be rather than have or the passive of sense 4 (as in they were sat down ‘they had sat down’ or ‘they were made to sit down’).
    ?1574 tr. H. Niclaes Exhortatio ix. f. 18 My beloued Children and thou Famelye of Loue: when yee are sat-downe at the Table to eate.
    1632 J. Hayward tr. G. F. Biondi Eromena 185 As soon as they were sitten down [It. assisi].

    They don’t seem to be implying that it’s been in continuous use (or at least, it isn’t clear that it has), and they also don’t make the same note for “to be stood”, which has an earliest citation from 1860:

    1860 Rep. Commissioners Corrupt Practices Gloucester Election 461 in Parl. Papers XXVII. 1 I was stood at the door, smoking a pipe along with a friend I knew from Birmingham.

    So they seem to doubt that this is a Venerable Survival. The quotation from the Anglo-Saxon Chronicle in the old edition was:

    sit 18b. In pa. pple. with is, was, etc. Cf. 23a (b). Now dial.
    c925 O.E. Chron. (Parker MS.) an. 922, Him cierde eall þæt folc to þe on Mercna lande ᵹeseten wæs.

    That quote is used in the revised edition under the form history for the past participle (or at least, something almost identical dated “anno 918”), but not under “to be sat”. I have no idea what’s up with that.

  62. David Eddyshaw says

    I’m not sure if “I’m sat” and “I’m sitting” contrast as stative versus dynamic as you suggest, but it’s an interesting thought. Worth pondering … I suspect you’re onto something with that.

    Kusaal (I couldn’t restrain myself any longer) body-position verbs, like “stand” and “sit”, are a very distinctive group which is neither fish nor fowl as far as this sort of thing goes. Morphologically, they all belong to the minority imperfective-only conjugation, other members of which are predicate-adjectival verbs like “be long”, “be bitter”, the verbs “to be”, and relationship verbs like “have”, “love”, “be better than.” However, unlike other one-aspect verbs they are dynamic, not stative, syntactically, and can make progressive/continuous forms. Diachronically, I suspect they are remnants of a stage when the language had separate imperfective-habitual, imperfective-progressive and imperfective-stative flexions; in the major conjugation the habitual and progressive flexions now coincide, and the stative has formally fallen together with the perfective.

    Not surprised if the “sat” construction doesn’t go back to OE; given that the use of “be” rather than “have” in perfects is still there in Shakespeare and the Authorised Version, you’d have thought that there would have been abundant instances of “is sat” and “is stood” in this sense from Early Modern English if the forms had been in continual use all that time. Thanks for looking it up.

    So I will revert to thinking that it is a newish construction, though nevertheless one well grounded in the behaviour of body-position verbs cross-linguistically.

  63. what about the hundreds of thousands of papyrological fragments sat in university libraries?

    Good point. More epicycles needed …

    I don’t see the problem. It’s perfectly normal to say “thousands of papyrological fragments sitting in university libraries” (at least I would say it without a second thought), and since “sat” = “sitting,” Bob’s your uncle.

  64. David Eddyshaw says

    Also a good point …

    However, it at least shows that you can’t just mechanically substitute “seated” for “sat” in these constructions, as I had been implying. And even leaving aside the thought of papyri sitting on their papyric backsides, this “sat” seems to have properties which rule out simply parsing it as a participle of any sort. I think it’s (perhaps) syntactically a derived deverbal adjective only usable predicatively; that would actually go quite nicely with with ktschwartz’s idea that the “sat” construction is stative rather than dynamic, and would account for the fact that you can actually constrain it to a habitual meaning in appropriate contexts: “I’m sat” is not simply equivalent to the progressive “I’m sitting.” There’s also the fact (if it is a fact*) that “sat” in this sense seems to require a complement, which is a thing some adjectives do, whereas “sitting” certainly does not.

    * I suspect that this may turn out to be yet another area where our native-speaker intuitions don’t all agree …

  65. David Eddyshaw says

    Developing ktschwartz’s idea (which I’m beginning to like a lot): “sat” in this sense would mean “in a state of sitting” (much as “asleep” is “in a state of sleeping.”)

    That gets round the papyrus rump difficulty and goes with the limitation to predicative uses (though it doesn’t encompass the need for a complement, it that is indeed a real thing.)

  66. I’m searching for [other] intransitive be-perfective verbs to compare with this “sat” and “stood”. Reminiscent of French, English seems to tolerate this for a few verbs of motion: “when I’m gone”; “Christ is risen”. Less relevant are other middle-voice things like “well-spoken” and “all partied out”.

  67. David Marjanović says

    I think this is an interesting consequence of set no longer being the straightforward causative of sit anymore (that’s preserved better in German), and likewise of stand having lost its causative (uh… “put up”) altogether. The merger of lie and lay is ongoing.

    Participles and adjectives with identical meanings but not quite identical usage are already a thing in Germanic and Slavic. I’m thinking of has died vs. is dead, which can’t be distinguished in French. Likewise there’s living vs. (a)live, constructed in German through desperate means: lebend vs. lebendig (adjective suffix tacked on the end of a complete participle).

  68. David Eddyshaw says

    In Kusaal, adjectives are quite freely derived from verbs, but only with habitual or stative meanings, so you can have e.g. pɛ’ɛlʋŋ “full” from pɛ’ɛl “fill” and kpiilʋŋ “dead” from kpi “die”, but there is no adjective meaning e.g. “seated”, because zin’i “sit/be sitting” is not stative (though there is a perfectly good adjective zin’idir, meaning “habitually connected with sitting”, e.g. kugzin’idir “stone for sitting on.”) In other word, Kusaal has the “live” type of adjective, but not “living”, and there are no “participles” in the SAE sense (the language uses clause catenation for that instead, in the misnamed “serial verb” constructions.)

    “Sat” and “stood” in the complained-of sense indeed resemble “full” rather than “filled.”

    English seems to tolerate this for a few verbs of motion: “when I’m gone”; “Christ is risen”

    “Christ is risen” is a formula borrowed from Early Modern English, when this was still normal for the perfect of intransitive verbs, but “when I’m gone” is an interesting one. (An odd man out among the Kusaal dynamic one-aspect verbs, which nearly all express body positions, is wa’e “travel”, but I think that is mere coincidence.) I suppose “not being there” is a kind of (anti-)state …

  69. David Eddyshaw says

    Reminds me of the Tathāgata them(not)self:

    https://en.wikipedia.org/wiki/Tath%C4%81gata

  70. you’d have thought that there would have been abundant instances of “is sat” and “is stood” in this sense from Early Modern English if the forms had been in continual use all that time

    Not only that, but the absence of is sat (except in the irrelevant causative sense in which babies are sat in high-chairs by adults) from North American English shows pretty clearly that it wasn’t around when anglophones were settling this continent.

  71. Lars Mathiesen (he/him/his) says

    FWIW, Danish has den er stillet/lagt/sat på hylden vs han har stillet/lagt/sat den på hylden. These are the participles belong to the causatives stille/lægge/sætte, the statives are stå/ligge/sidde with participles stået/ligget/siddet. (I suspect suspicious things when the causatives have stronger forms than the base verbs, but there it is).

    (Which of the verbs to use for placing a specific item on a shelf is a matter of pragmatics).

    (Danish has kept the be-perfect for non-stative intransitives, but extended have-perfects to statives: den har ligget på hylden = ‘it’s been lying on the shelf’ [and maybe it still is, but that’s why it’s dusty]).

    Also the “proper” causative of stå seems to have evaporated; I don’t have the reconstructive nous to posit a possible PG form and carry it forward. How about stænde, though? In any case, stille is derived from an LG adjective, akin to E still; I think the causative-of-stand sense was formed in and borrowed from MLG, while Danish (like E) formed another stille = ‘make still’ on its own. (Also stilne = ‘abate’ of winds and the like).

    Lastly, har døet is rare, because when is having died (at have døet) different from being dead (at være død)? (ODS takes the position that død is “simply” an irregular way of spelling [and pronouncing] the participle — except in the have-perfect and as a phrasal verb: vinden er døet hen; note be-perfective because change of state). But it’s perfectly cromulent in counterfactuals: jeg kunne have døet hvis der ikke fandtes penicillin — though even there I think current practice would prefer jeg kunne være død nu […]. Adverbs FTW! But I expect this particular verb to exhibit pragmatic lacunae in its conjugation in most languages, because semantics.

    (That’s disregarding the jocular use of transitive constructions with verbs that are normatively strictly intransitive, like and : han blev gået or de har gået ham = ‘he was fired/they fired him’; even at gå nogen = ‘to fire someone’. You can supply your own semantics for at dø en mand. The point is that these can’t be mistaken for any intransitive use so you are forced to parse them as wilfully non-standard).

  72. Lars Mathiesen (he/him/his) says

    (Edit timer ran out before I ran out of stuff to run off about: I think Sw jag kunde ha dött is more current than the Danish equivalent, FWIW).

    And I got to thinking if død is really that irregular, maybe it’s just a case of being the only verb in its aspect class ending in a stressed vowel. But no, has much the same properties and we don’t say ‡jeg er gåd.,

  73. “He was sat there” is often used colloquially in the UK to mean “He was sitting there”. It still sounds wrong to me, but I suspect common usage will dictate when “was sat” is acceptable. On the other hand, “he was stood up” is ambiguous; it could mean either that he was standing up (intransitive) or that his date didn’t turn up (transitive).

    Sorry – I’m late to the party but another explanation of “codes that we could read but aren’t” could mean “codes that we could read but aren’t [actually codes] or [actually reading].”

    “Could” implies possibility; “aren’t” implies actuality and the two are juxtaposed rather oddly. It’s just not a well-constructed, unambiguous sentence and needs a good copy editor.

    I’ve often wondered if a shorthand system based on sound such as Pitmans could be adapted to become a writing system for all languages.

  74. David Eddyshaw says

    This question of how derived adjectives relate to cognate verbs aspectually is quite interesting in its own right, I think. Obviously it’s highly language-dependent, but it would be nice to have some sort of neat framework of categories to classify the various possibilities with.

    I suspect suspicious things when the causatives have stronger forms than the base verbs, but there it is

    Kusaal makes inchoatives out of body-position verbs (e.g. “sit down” from “sit/be sitting”) with a derivational suffix -n which not only appears nowhere else in derivation but is the only derivational suffix which regularly alters the tones of the root it’s attached to. Body-position verbs are weird. (Hey! A robust cross-linguistic typological generalisation! Call me Greenberg!)

  75. David Eddyshaw says

    The trouble with CGEL is there is so damn much of it that you almost need to know the answer beforehand in order to be able to look something up in it. I see, however, that it has an extended treatment of what it calls “adjectival passives” in Chapter 16, 10.1.3 (pp 1436 ff.) The section proposes various distinguishing tests, none of which is a lot of help with “sat/stood” as far as I can see, but it does mention (p1440) “Adjectival passives with specialised senses”, where

    There are a few adjectives that are morphologically related to the past participles of verbs but whose meanings have changed, so that they are no longer comparable to verbal passives with the same forms, and their connection with passives proper is purely historical:

    She’s bound to win. We’re engaged (to be married.) Aren’t you meant to be working on your assignment? His days are numbered. Are you related? I’m supposed to pay for it. He isn’t used to hard work.

    I reckon “sat” and “stood” (in the uses under discussion) belong here (though it seems a bit misleading to call these “passives” at all, really: their historical origin is not relevant to current usage.)

    It seems likely to me that OED’s citing of the passage from the Anglo-Saxon Chronicle is misleading (well, wrong, in fact.) This is not actually the same usage as the Old English one, which was simply a (perfectly regular) perfect used resultatively à la Swahili. The construction under discussion is indeed a Horrid Innovation (worse yet, a dialect form impertinently encroaching on the sacred profanity-free Standard, and not a survival from the Good Old Days at all. The cheek of it!)

  76. “belong here”

    Why?
    These examples are markedly different in that they allow very straightforward passive interpretation (except maybe “used to”: is it from the noun “use” in the older sense “habit”?). Someone obligavit you, and you’re bound…

    Meanwhile there are fallen and risen and drunk(en)…

  77. And some of us are even woke….

  78. Also done work

  79. David Eddyshaw says

    These examples are markedly different in that they allow very straightforward passive interpretation

    Not in the least; less so than with “sat” or “stood”, if anything.

    “She’s bound to win.” Bound by whom? And can you bind someone to win?
    “Are you related?” So who related you? (With this ring I thee relate …)

    I think you’re confusing diachronic origin with synchronic usage. (Hey, we’ve all been there …)

    “Bound” (in this sense, “inevitably destined”) also shares with “sat”/”stood” the property of being predicative-only, I notice, and requiring a complement. Similarly with “supposed” and “meant” in the sense “morally obliged.”

  80. David Eddyshaw says

    (Though you can say “It’s all meant” if you’re of a Panglossian/Pollyanna-ish disposition.)

  81. David Eddyshaw says

    fallen and risen and drunk(en)

    “Fallen” and “risen” are simply past participles, usable as adjectives Because Participle. Neither can be passive at all, because “fall” and “rise” are intransitive.

    “Drunk” is just another case of an etymologically verbal form acquiring a new life as an independent adjective. If I’m drunk, it does not follow that somebody drank me, and there is no way of making “I’m drunk” into a passive. (Except in The Hitchhiker’s Guide to the Galaxy: “It’s unpleasantly like being drunk.”)

  82. Neither can be passive at all, because “fall” and “rise” are intransitive.

    Yes, exactly. But isn’t it what we need?

  83. @DE, in Russian obliged is обязан < об- -вязан (around-tied). An obvious passive form and an obvious calque.

    I thought English also calqued it, and the question "by whom" must be addressed to Latins. Also the point of passives is making the noun a patient rather than specifying the agent…

    But now I disagree with what I wrote. English “bound to” (as in bound to win) does not feel like an extension of “obliged -> must”, is not identical to anything Russian and I’m not sure about French. Maybe it is “bound” in the sense “ready”, as in “eastbound”. Interference from bound “obliged” could add the sense of necessity…

  84. “I think you’re confusing diachronic origin with synchronic usage. ” – I thought that they all are sychronously felt as passives, if it’s what you mean by synchronous usage. I also thought they are literary borrowings (except “meant”, and “bound” )…

  85. David Eddyshaw says

    No, to both points. Not felt as passives; not literary borrowings. Five are ultimately from French or Latin, but so is a great deal of everyday colloquial English vocabulary (“use”, “very”, “simple” … “language” …)

    This is a language which has borrowed personal pronouns

  86. I thought that they all are sychronously felt as passives

    No, as DE said.

  87. PlasticPaddy says

    Compare “he was bound to lose (concede defeat)” with “he was forced…”. For me the formulation with “forced” implies a more specific external cause or (probably consciously acting) agent, even though the objective reality and the speaker’s intention could be identical in the two cases. Re sat/sitting, can you say “I was sat at the bar, minding my own business / waiting for a friend”? In Irish English I think you can have double participles, e.g., I was standing waiting on a bus.

  88. David Eddyshaw says

    Indeed: if I say “I’m forced to do that”, you could reasonably ask “Who’s forcing you?”; it’s just a passive.
    But if I say “I’m bound to do that”, it makes no sense to ask “Who’s binding you?”

    can you say “I was sat at the bar, minding my own business / waiting for a friend”?

    Sure. It has no bearing on whether “sat” is actually functioning as a participle there, of course: you can say with equal grammatical propriety “I was miserable, drowning my sorrows at the bar” or “I was drinking Guinness, drowning my sorrows at the bar.”

  89. Я стоял ждал автобуса…

  90. Indeed: if I say “I’m forced to do that”, you could reasonably ask “Who’s forcing you?”; it’s just a passive.

    @PP, DE, but passives do not require agents:-/ English is full of impersonal passives…. As I understand it is not a problem for DE, but DE implies that for each impersonal passive construction there exists a verb (in its other forms), and moreover, each user of an impersonal passive construction believes that naming the agent is in principle possible even if she does not know him.
    This looks like an abstract logical requirement:/

  91. You don’t know who are “they” in “they say”, you don’t know what is “it” in “it’s raining”, but you can’t claim that “say” does not have verbal semantics and is not a “a verb”.

    Of course, “the rumor is” equals to “they say” at some level (and “would you like a pint of Guinness?” equals to silently pointing at a barrel), but not at other levels.

  92. This looks like an abstract logical requirement:/

    It sounds to me like it is you who are being abstractly logical; to a native speaker of English, these do not feel like passives.

  93. About literary borrowings: Book worms converse too… we even reproduce.

    I didn’t mean anything highfalutin’ like “words that a literate schoolboy won’t use in an informal conversation with his classmates”. Rather “words and expressions that an illiterate speaker would not use some time ago”.

    In modern Russian there is a stratum of words and expressions that I expect to hear from the later, and a stratum that I expect to hear from the former.

    E.g. the usual word for “a relative” is родственник, where род- is the root and -ственник is a pile of suffixes. The word is old enough (17th century?) but I suspect that this pile of suffixes (one of them abstract/collective) was produced by literate people and modelled after Slavonic words. It is long enough for me to use it less often than I would use a shorter word. And also cold (not warm) and formal enough, which means I associate it with more formal contexts.

  94. “It sounds to me like it is you who are being abstractly logical; to a native speaker of English, these do not feel like passives.”

    @LH, but I was commenting on DE’s (logical) test.

    I was not telling that you (or DE) were wrong:/ DE supported his interpretation with an argument and I am surprised by this argument.

  95. David Eddyshaw says

    English is full of impersonal passives

    If you mean, passives with no agent expressed, sure. Indeed, that’s often the whole point of using them. (It’s incorrect to call this “impersonal”; it’s quite hard to come up with cases where a passive subject actually is impersonal in English, although it can be done: “It’s said to rain very little in the Sahara.” It’s definitely far from common. Latin and Welsh are better at it …)

    But it makes no sense whatsoever to call a construction “passive” if there is no corresponding active construction. Perhaps you would call “I’m bound to …” constructions “deponent”?

    Rather “words and expressions that an illiterate speaker would not use some time ago”

    These expressions just don’t fall into that category in English. They really don’t. They are not literary in the slightest.

  96. DE: “It seems likely to me that OED’s citing of the passage from the Anglo-Saxon Chronicle is misleading (well, wrong, in fact.) This is not actually the same usage as the Old English one, which was simply a (perfectly regular) perfect used resultatively à la Swahili.”

    That was the old OED (from 1911) that cited it; if it’s unambiguously a perfect, that explains why they dropped that quotation in the revision. They also dropped another early quotation:

    c1290 St. Brendan 281 Þo heo weren alle i-sete Þare cam on and seruede.

    … presumably for the same reason. Also, the old OED didn’t give any grammatical interpretation to their sense 18b; they just identify the form, “In pa. pple. with is, was, etc.”, without classifying it as passive, perfect, or whatever. They also didn’t give a definition, implying it meant the same as the preceding 18a, “To seat oneself; to take a seat; to sit down”—and that, I think, implies in turn that all the quotations must have been interpreted as perfects in 1911, since they were read as actions. (Burchfield (1996) and Butterfield (2015) apparently overlooked the be-perfect interpretation.) The revision re-interprets the remaining Middle and Early Modern English quotations as ambiguous and possibly continuous with the present-day “is sat”.

    If that’s what they were thinking in 1911, then they were mistaken in including this Lancashire dialect quotation in the same category:

    1864 J. Ramsbottom Phases of Distress 12 At th’ eend o’ th’ day..aw’m sat at whoam.

    … since that *does* seem to be the same as the present-day “is sat”, especially with more context:

    At th’ eend o’ th’ day, mi wark o’ done,
    An’ quite content, aw’m sat at whoam;
    Mi childher brimmin o’er wi fun,
    Ull singin’ reawnd abeawt me come.

    (EDIT to add:) So I wonder if this “is sat” was still regional/rare/new/stigmatized enough in 1911 that it wasn’t recognized yet in Oxford as a separate thing from the archaic perfect. They also didn’t notice “is stood” at the time.

  97. David Eddyshaw says

    Thanks, kt.

    The spelling “childher” is odd. I wonder what it is intended to represent?

    It must be difficult to impossible in texts from the period when the be-perfect was still in use to differentiate between an ordinary perfect used resultatively and this new-fangled sort of “is sat.” In fact, I’m not even sure that there is meaningful distinction: the “sat” construction is perhaps only identifiable as such at all after the demise of be-perfects. It does make one wonder again about whether the construction was – originally – a survival.

  98. David Eddyshaw says

    Interestingly, such an origin would be impossible for any of the examples that CGEL gives.

    Mind you, that doesn’t mean I can’t lump this “sat” and this “stood” in with “bound” etc as being, synchronically, the same sort of animal: it would just mean that the various constructions in question had arrived at this point (viz looking like passives, but not actually being passives) via different pathways: in CGEL’s examples, by a meaning shift from an original past participle into something which can no longer by analysed that way, and in the “sat/stood” case by reinterpretation of an old active intransitive perfect as a “passive.”

    And interpretation of this “sat” and this “stood” as (predicative-only) adjectives syntactically distinct from the homophonous past participles seems to fit very nicely with the actual usage, and the stative quality that you pointed out.

  99. Jen in Edinburgh says

    Someone who knows more of the technical language than me might want to have a look at https://dsl.ac.uk/entry/dost/sit_v

    p.t. and p.p. forms in set, etc. = sat, seated, without causal implication are included in Sit v.’, it says.

    Is this the kind we’re looking for?

    1617 Misc. Abbotsf. C. 311.
    This witt was sett smoaking … in ane obscure chymney corner at Leethe

  100. David Eddyshaw says

    Interesting stuff! DM alluded to this set/sit business in passing above, as well. I hadn’t quite grasped what he was driving at before.

    So: another potential source of the condemned “sat” construction: from an actual passive of “set” reinterpreted as a form of “sit.” (Some of the OED’s premodern examples look rather as if they might belong to “set” rather than “sit”, too.)

    It wouldn’t explain the “stood” version, though analogy with “sit” might have been involved there too. On the other hand, “stand” itself is used as the causative of intransitive “stand” in modern English; I can’t remember (if I ever knew) what the equivalent in older English was, but if there was a distinct causative verb I imagine that it too would have been liable to get confused with the intransitive simplex verb.

    [It’s all a lot simpler and clearer in Kusaal: zin’i “be sitting”, zin’in “sit down”, zin’il “sit someone down”; zi’e “be standing”, zi’en “stand up”, zi’el “stand someone/something up.”]

  101. I can’t say what the spelling “childher” represents, but the author is very consistent in replacing t/d with th/dh before r or -er. Just flipping through the first few pages, I noticed: sthrivin, sthrife, sthrings, sthrippin’, sthreams, sthraw, counthry, betther, facthory, thrials, disthress; dhrops, dhrives, dhress, dhresser, dhreadful, wandhert, wondhrous, cindher. Could it be what phoneticists call “anticipatory retroflexion”? Or affrication of t and d before r? Is that a thing in Lancashire dialect?

  102. David Eddyshaw says

    Affrication of the stop in /str/ seems weird, though I suppose anything is possible …

    I wonder whether (per contra) the spelling might actually represent dental rather than alveolar stops? Traditional Irish English does (contrary to the stereotype) distinguish the original pairs t/θ d/ð, but the original fricatives have become dental stops. That might have provided a prototype for using the graphs th/dh for dentals.

    https://en.wikipedia.org/wiki/Th-stopping

  103. PlasticPaddy says

    Father Jack: “Dhrink!”

  104. David Eddyshaw says

    The WP page on th-stopping links to an interesting if vaguely depressing paper about what sort of foreign-learner errors (both of pronunciation and grammar) particularly irritate L1 English speakers, and/or lead to actual misunderstanding; much of it is interestingly counterintuitive. It makes the point that the strategy of adopting renderings of difficult phonemes like /θ ð/ as stops or as /f v/, which some L1 speakers themselves do, may actually run afoul of the fact that such pronunciations are stigmatised among native speakers. And

    There is even some indication that some pronunciations associated with Ireland or Scotland (see 4.6) were not judged more leniently by respondents from those areas

    I think people can have quite firm ideas of what foreigners ought to sound like, which do not always coincide with talking exactly the way that they themselves do (I mentioned elsewhere that I was myself on the receiving end of this attitude with Hausa.)

    A surprising amount of politics; in particular, pointing out that the facts underlying Jennifer Jenkins’ concept that native speakers don’t own English and have no title to control “International English” are not always all that factual.

    https://www.theguardian.com/education/2004/jan/22/tefl.wordsandlanguage

    From van den Doel’s paper I glean the exciting factoid that, while Dutch learners of English mostly believe that UK speakers are more critical of the learners’ English than Americans are, the reality is actually strongly the other way round.

    https://dspace.library.uu.nl/bitstream/handle/1874/13381/Doel-13-completetext.pdf

  105. Heh, the system here reported back first that I can no longer edit my rather long comment (after just a couple of minutes). Then it failed to show my comment, but adjured me not to duplicate it!

    Well, now I add something that OED now says about its set entries:
    • For many years the verb to set has been cited as the longest entry in the OED. But a recheck shows that it has at last been toppled from this position. The longest entry in the revised matter is represented by the verb to make (published in June 2000). However, it is quite possible that set will regain its long-held position at the top of the league of long words when it comes itself to be revised.
    • In ranking order, the longest entries currently in the online Third Edition of the OED are: make (verb – revised), set (verb), run (verb), take (verb), go (verb), pre- (revised), non- (revised), over- (revised), stand (verb), red, and then point (the noun – revised).

    Now I’ll attempt to post my original comment within this one. LH might remove duplicates:

    David Marjanović said:
    • I think this is an interesting consequence of set no longer being the straightforward causative of sit anymore (that’s preserved better in German), and likewise of stand having lost its causative (uh… “put up”) altogether. The merger of lie and lay is ongoing.

    Jen in Edinburgh cited a site:
    • “p.t. and p.p. forms in set, etc. = sat, seated, without causal implication are included in Sit v.”

    The Guinness Book of Records (“in cervisia veritas”) weighs in:
    • “The word with the most meanings in English is the verb ‘set’, with 430 senses listed in the Second Edition of the Oxford English Dictionary, published in 1989. The word commands the longest entry in the dictionary at 60,000 words, or 326,000 characters.”

    David Eddyshaw remarked, before moving on to analysis:
    • Interesting stuff!

    Yes, and a nest of etymological tangles. It is well to set out more of the evidence.

    The first entry for set as a verb in OED gives this etymology:
    • Common Germanic: Old English sęttan = Old Frisian setta (modern Frisian sette), Old Saxon settian (Middle Dutch, Middle Low German setten , Dutch zetten), Old High German sezzan beside sazzan (Middle High German sezzen, German setzen), Old Norse setja (Swedish satta, Danish sætte), Gothic satjan; causative of *setjan (sitjan) to SIT v.

    Followed by this highly relevant note (pay no attention, David E):
    • Confusion between set and sit arose as early as the beginning of the 14th cent., owing partly to the identity or close similarity of the forms of their past tenses and past participles, and partly to the identity of meaning in some uses, as between to be set (= seated) and to sit; compare SIT v. (etymological note and 5a note). For cases of mere substitution of forms of sit for forms of set, see 1 γ. , 2 ζ. forms. The spelling sett is still sometimes found in technical senses; compare SET n.1

    That OED entry gives this broad meaning first, again followed by a useful note:
    • I. To cause to sit, seat; to be seated, sit.
    The intransitive sense ‘to sit’ (5) was apparently developed out of the reflexive and passive uses of the original transitive sense of ‘to seat’. Set, being thus used synonymously with sit, became capable of taking its other senses and constructions (see 5d, 5e, 6, 7).

    Earlier I had said, before we spiralled even further from the original topic, in time-honoured Hattic tradition:
    • To explore further I would substitute a regular verb for read, and try for a convincing parallel that does without the modal ambiguities of could, issues with that, forms of to be that are not common to we and to they [the codes], and so on.
    And I stand by that principle. In the present case (set, sit, seat v., stand, etc.) progress is bound to be held back because we can’t readily simplify anything with parallels. The complexity is baked in.

    Perhaps there are mix-ups involving passivity. Many French constructions with auxiliary être don’t feel at all passive: “Aujourd’hui, maman est morte. Ou peut-être hier, je ne sais pas” (Camus); “Je suis venu”; “Nous sommes descendus du train à Marseille”. No agent “meurt maman” (“dies my mother”), or “me va” (“comes me”), or “me descend du train” (“gets me off the train”). But of course, many other constructions are passive: “Nous étions assis pour dîner a six heures” (“We were seated for dinner at six”). There is likely to have been an agential host, who reports: “Je les ai assis pour dîner a six heures” (“I sat them down for dinner at six”). Yes, asseoir works like that. The reflexive gives instances of its working like that: “Nous nous sommes assis pour dîner a six heures” (“We sat down to dinner at six”; “We sat ourselves down for dinner at six”).
    Now, in English too “we were sat” need not suggest passivity, just as a good number of Googled forms (mostly pre-20C) don’t suggest it: “Here I am arrived at Hauraki”; “And when so general an odium was risen against them …”; and Tennyson’s “I am become a name”, in “Ulysses”.

    Finally, for what it’s worth to seat (as in “we were seated for dinner at six”, with either passive or non-passive meaning) is quite a latecomer. OED’s first example, in all the many meanings at “seat, v.”:
    • 1577 B. GOOGE tr. C. Heresbach Foure Bks. Husbandry IV. f. 172 Euery house is not so seated, as it hath earable ground about it.
    (We love earable, yes? Recalling Shakespeare’s “uneared womb”.)

  106. Correction:

    or “me va” (“comes me”) > or “me vient” (“comes me”)

    Addition:

    Shakespeare prefers to be with arrive:

    BRUTUS. Hark, he is arrived.
    March gently on to meet him.

    MALCOLM. I would the friends we miss were safe arrived.

    Just as with become and return:

    PHILO. … And is become the bellows and the fan
    To cool a gipsy’s lust.

    BENEDICK. Look! Don Pedro is returned to seek you.

    But he has also to have:

    LUCENTIO. And happily I have arrived at the last
    Unto the wished haven of my bliss.

  107. David Eddyshaw says

    Perhaps there are mix-ups involving passivity

    Well, yes. That’s what I’ve been saying.

    French (now you mention it) uses être instead of avoir to make the compound tenses of reflexive verbs and a few (extremely common) others, mostly verbs of motion, but also including naître and mourir; maman est morte is active passé composé and not passive at all:

    https://en.wikipedia.org/wiki/Pass%C3%A9_compos%C3%A9#Auxiliary_%C3%AAtre

    This is not dissimilar to the rules for older English and Germanic in general; the periphrastic perfect seems to be a Western European Sprachbund thing. I blame the Romans (and possibly the Greeks.)

    It’s pretty clear that, regardless of its actual origin, “I’m sat at the table” (in the sense we’re discussing) is not passive. It may conceivably have arisen from a passive historically, like “I’m bound to …”; this is where the sit/set confusion could be relevant. Or it may have arisen from a perfect with the auxiliary “be”, as the older OED entry seems to imply; or it could be just an independent new formation leveraging the cross-linguistic weirdness of body-position verbs. None of this affects the analysis of the contemporary construction, interesting though it is in itself.

    (pay no attention, David E)

    Hah! I see through your transparent reverse psychology. I won’t pay attention. So there!

  108. I won’t pay attention. So there!

    (I’ll ignore that.)

    Well, yes. That’s what I’ve been saying.

    It would save us a great deal of research effort if you could tell us at the time what you were saying. Still, better lace than leather.

    This is not dissimilar to the rules for older English and Germanic in general; the periphrastic perfect seems to be a Western European Sprachbund thing. I blame the Romans (and possibly the Greeks.)

    I blame the Croatians, and a whole wardrobe-shelf-full of odd Slavs:

    Nacrtao sam krug.
    drew [perfective] [I] am [a] circle
    I drew a circle.

  109. David Eddyshaw says

    It would save us a great deal of research effort if you could tell us at the time what you were saying

    https://languagehat.com/ancient-greek-shorthand/#comment-4493796

    where I vouchsafed: “the OED is surely wrong in calling this ‘passive’: it actually contrasts with the passive uses.”

    What can I say? I lack the gift of lucid exposition. In my defence, this is due to years of deliberate obfuscation in my professional capacity. (“Bedside manner” is the technical term.)

  110. What can I say? I lack the gift of lucid exposition. In my defence, this is due to years of deliberate obfuscation in my professional capacity. (“Bedside manner” is the technical term.)

    Running that through Google Translate, um … as we speak.

    Correction:

    drew > drawn

    Nacrtao sam krug.
    drawn [perfective] [I] am [a] circle
    I drew a circle.

    Of course, others here would remind us that Bulgarian does it without the auxiliary:

    Начертах кръг.

    (Along with a side-table-drawer-full assortment of other such tongues.)

  111. DE, you write lucidly. I don’t know why Noetica was harshing you.

  112. Noetica: “I add something that OED now says about its set entries … it has at last been toppled …”

    Naturally, Language Hat posted about that when it happened: “SET” UPSET. (And you were there!)

  113. I don’t know why … was harshing you.

    Seconded. Clearly there was an exploration going on, then a hypothesis, then a search for supporting and counter- examples.

    ‘Bedside manner’ my foot! Patients might not follow the jargon, but they know when they’re being patronised. DE was not patronising with sat/stood, but thinking out loud.

  114. ktschwarz:

    Ah yes! I must get my memory assessed. I was right in the thick of that discussion of set fifteen years ago, with my single comment: “Yes, the OED spelling is tit-bit or tid-bit, with a hyphen. …” I grow old, as Hat is wont to say of himself. Anyway, it’s good that you have now updated the former thread with OED’s current text about record-holding entries.

    Y:

    DE, you write lucidly. I don’t know why Noetica was harshing you.

    AntC:

    Seconded.

    (Settling a pillow by my head:) That is not it at all. I have observed that David E likes a bit of a fun-and-bustle rollick around the threads, so I continued certain themes he had begun, such as “I lack your human concept of ‘relevance’.” I join him in wondering at the ways of you humans, so I ran with the joke. And yes: he writes lucidly. But there were subtleties, twists, caracoles, the odd reversal, and a slew of interstitia; so it was hard to discern what the current message was. For me it was, en tout cas. (Did I say? I grow old …)

  115. I’m all for a bit of plain English. Say what you’ve got to say, shut up and sit down. After umpteen lengthy posts nobody seems to agree on anything, or if they do, they take so long to say so that my tiny brain can’t work out what they actually are saying. Take that as a failure on my part rather than on yours.

    So, to help me, please answer with a simple “yes” or “no”: Is the sentence “There are codes that we could read but aren’t” (1) grammatical (2) unambiguous (3) simply colloquial? No explanation required, just simple one-word replies to numbers (1) to (3) (if possible).

    Thank you for your indulgence.

  116. David Eddyshaw says

    Yes! Yes! O YES!

    (YMMV.)

  117. 1. Sometimes.
    2. No.
    3. No.

  118. January First-of-May says

    I’m all for a bit of plain English. Say what you’ve got to say, shut up and sit down.

    For what it’s worth, I do have to admit that Noetica’s posts tend to stray quite a bit farther from the plain side than usual on LH. Which has its own charm, of course, but it does make it noticeably harder to figure out the actual intended point sometimes.

    (This might have contributed to my apparently-mistaken perception of them as female. Though the last letter of the username was probably more important.)

    So, to help me, please answer with a simple “yes” or “no”

    1) Yes, but I’m not sure if it’s grammatical in the intended meaning
    2) Technically no, practically probably yes
    3) I wouldn’t say “simply”…

  119. David E: “well grounded in the behaviour of body-position verbs cross-linguistically”

    Hmm, but Northern England English *doesn’t* do this with body-position verbs generally, does it? That is, these aren’t possible, are they?

    I’m laid on a beach towel enjoying the sun.
    He’s leaned against the wall smoking a cigarette.

    At least, I’d think that if these were possible, they would have been brought up by now.

    Trying to think of other body-position verbs… let’s try sprawled. (Originally a body-position verb in Old and Middle English, according to the OED; other uses are later extensions.) Sprawled passes all of CGEL’s tests for “adjectival passive” (thanks for the pointer!), i.e. it’s an adjective, passive only by historical derivation. (There’s no active and transitive counterpart to e.g. A familiar, peaceful scene was sprawled out in front of him, though it could be reflexive, sprawled itself out.) It can even appear in attributive position: they stepped over the sprawled bodies.)

    CGEL also notes that such adjectives always have a stative meaning (they may be used with change-of-state verbs like become, but then it’s the verb that’s dynamic, while the adjective describes the resulting state). And yes, sprawled describes the state resulting from the action of sprawling.

    So if sprawled (and likewise stretched) made the transition to adjectives in standard English, why not sat? Still puzzling over that.

  120. Let’s try sat and stood with CGEL’s tests for “adjectival passives”. They fail two of the tests—very stood and unstood are impossible (right?)—but those conditions are sufficient, but not necessary. They *do* pass the one test that they say is necessary, the use with remain and similar verbs (I have no intuition for this, but Google provides examples; native speakers please check):

    So the builder remained sat on his bar seat when Fred introduced the girl to him (Tottenham)
    we remained sat on the runway awaiting a slot to return to London (UK)
    You could have remained sat, strapped in place, tears streaming down your face whilst the staff attempt to pry you from your paralysed position (UK)
    leaving Havertz on while Pierre-Emerick Aubameyang remained sat on the bench. (Evening Standard, UK)

    he remained stood with his back to us all the time we were in the cafe (TripAdvisor, Newark-on-Trent, England)
    the female police officer remained stood at the rear of the police van. (Scotland)
    their stars remained stood while Barnsley’s players did continue to take the knee (The Sun, UK)
    Fans remained stood despite the continuing rain and no play during India vs New Zealand (picture caption on shutterstock; match at Nottingham, England)
    A group of Christian leaders then prayed over Donald Trump, heavily praising the US President, while the majority of the crowd remained stood. (Irish site)
    For the next half hour there was a stand-off. We remained stood in the middle of the road (Liverpool)

    Granted, all of CGEL’s examples are from verbs that are definitely transitive, while sit and stand most often aren’t. But we could speculate that sat and stood are adjectives that developed from the transitive uses of the verbs.

  121. I’m not sure I’m following all of kts’s drift but either of

    I’m laid/sprawled on a beach towel enjoying the sun.

    seem unexceptional. (And handily avoids me having to stress about ‘laying’ vs ‘lying’. One of those is for cutlery on the picnic-cloth beside the beach towel.)

  122. David Eddyshaw says

    But we could speculate that sat and stood are adjectives that developed from the transitive uses of the verbs

    Yes, that’s why Jen and DM’s points about sit/set confusion seemed pertinent.

    It seems more likely on first principles than isolated survival of be-perfects; apart from anything else, I would have thought that a be-perfect-in-resultative-sense origin would have given rise to e.g. “I’m sat down at the table” and “I’m stood up at the bar.” Those are both possible, but the “down” and “up” seem fully dispensible-with.

    I’m laid on a beach towel enjoying the sun.
    He’s leaned against the wall smoking a cigarette.

    Actually, yes: both of those are grammatical for me. [EDIT: ninja’d by AntC.]

    Mind you, what counts as a body-position verb is pretty language-dependent: Kusaal has, for example, separate one-aspect verbs for (at least) “sit”, “stand”, “lie prone”, “lie supine”, “kneel”, “squat”, “have the neck extended”, “have the neck bent”, “hide behind something”, “lean” (of a person), “lean” (of a thing), “hang”; “be stuck to” behaves like these verbs too, both morphologically and syntactically. “Be awake” is a one-aspect verb, too: I suppose you could construe that as a body-position verb if you squint hard enough.

  123. remained stood/sat

    [Native speaker here, about a third Yorkshire]

    Those examples I’d call acceptable but a bit jarring: ‘remain’ is formal register; ‘stayed sat’ sounds more natural.

  124. David Eddyshaw says

    In fact, thinking about it, the acceptablity of “I’m laid on a beach towel enjoying the sun” supports the idea that “I’m sat at the table” is a repurposed* passive of a causative “sit.” There is, similarly, no implication at all that some external agent laid me on the beach: I bear sole responsibility for it myself. Formally, however, the construction is indistinguishable from a passive.

    * I think “repurposed” is the key point in this. Ultimately, I don’t think anything much turns on whether you call this a syntactic innovation (so it isn’t actually “passive” any more) or just an extension of the semantic role of a body-position causative-passive to cover this kind of “deponent” use where no external agent is actually implied at all.

  125. David Eddyshaw says

    Interestingly, you can’t use the inchoative derivative of a Kusaal body-position verb in the perfective with a stative meaning.

    You can say

    O kpi nɛ.
    he die STATIVE
    “He’s dead.”

    interpreting the confusingly multipurpose particle as a stative marker, which is its default meaning when it immediately follows a verb perfective.

    But it is not possible to take in this sense after digin “lie down”:

    O digin nɛ.
    he lie.down FOCUS

    This is precisely because, in Kusaal, body positions aren’t “states.”
    Instead, the has to be taken as marking constituent focus: He’s lain down.

    My informants, who were very good at thinking up imaginary contexts where you could say something that at first blush seems ungrammatical, suggested: “Someone calls at your house; he thinks you’re out but I’m explaining that you’ve gone to bed”; and “You’ve said: the child looks filthy. I’m replying: He’s been lying down.”

    an extension of the semantic role of a body-position causative-passive to cover this kind of “deponent” use where no external agent is actually implied at all

    Which is what drasvi was driving at, now I think of it. Apologies, drasvi!

  126. “who were very good at thinking up imaginary contexts where you could say something that at first blush seems ungrammatical,”

    After having seen many essays by Russian learners and corrections by Russian speakers, I would say, we are good at inventing interpretations that make absolutely perfect Russian ungrammatical.

  127. @DE, I absolutely spoke about semantical passivity.

    But I said that constructions like “you are supposed to…” (but not “bound to”) are different in that they obviously have a passive interpetation.

  128. David Eddyshaw says

    Thinking about “I’m laid on a beach towel enjoying the sun” even more …

    There actually is a problem with interpreting this as passive, even leaving aside the question of agents.
    The problem is aspect.

    If it’s taken as passive, then it’s not continuous/progressive aspectually (that would be “I’m being laid on a beach towel”, which is actually unambiguously passive.) If you interpret it as passive, it has to have a habitual/timeless meaning:

    “I’m laid on a beach towel [by my trusty retainer] every Wednesday.”

    If you add adjuncts to force a “just this present moment” interpretation, it seems to make the passive reading awkard-to-impossible:

    “I am laid on a beach towel at present, reading a book.”
    * “I am laid on a beach towel [by my trusty retainer] at present, reading a book.”

    This (I think) is what ktschwartz was onto in saying that the “I’m sat” construction has a stative quality. It’s actually, like “I am suntanned”, aspectually neutral, and compatible with either a just-at-present reading or a timeless one: the distinction that in dynamic verbs is made by contrasting the “present continuous” with the “simple present.”

    So: if taken as passive, the form distinguishes

    “I’m laid on a beach towel [by my trusty retainer] every Wednesday.” – habitual/timeless

    “I’m being laid on a beach towel [by my trusty retainer].” – continuous/progressive

    In other words, the passive construction is dynamic, and changes for aspect.
    But in the complained-of anti-Soanesian sense, there is no such distinction:

    “Every Wednesday, I’m laid on a beach towel in Hartlepool thinking about transitivity.”

    “At this very moment that I’m texting you, I’m laid on a beach towel wishing you were here so that we could talk about transitivity together like we used to.”

    So the construction does not distinguish simple from continuous present.
    Tipped off by this, I found CGEL p1431 on “Verbal passives and adjectival passives”

    The verb be serves, of course, not only as a passive marker but also as a copula, taking a predicative complement in the complex-intransitive construction. And there can be ambiguity between the two constructions:

    The vase was broken. [be-passive or complex-intransitive]

    As a passive (of the short variety) this describes an event, as does the active clause Someone broke the vase. As a complex-intransitive clause it describes a state – the state resulting from someone or something breaking the vase. In the first interpretation broken is a bare passive clause consisting of just the head, the verb broken; in the second, broken is an adjective. We will follow the widespread practice of describing broken in the second sense as an adjectival passive, but it is to be emphasised that this represents an extended and derivative sense of the term ‘passive.’

    So I think that what is going on with the “I’m sat” construction is that a straightforward verbal passive (of a position verb used causatively) has been reanalysed as an adjectival so-called ‘passive.’ (This is just a wordier version of ktschwartz’s original insight.)

    Relevant to the sit/sat thing is CGEL p307

    With a relatively small number of verbs the intransitive denotes a state and the transitive the bringing about of that state: I leant the ladder against the wall. This is found with verbs of position, such as hang, rest, sit, stand, and a few others …

  129. But I said that constructions like “you are supposed to…” (but not “bound to”) are different in that they obviously have a passive interpretation.

    The interesting thing there is that the passive interpretation implies a completely different meaning: “You are supposed [by someone] to do X” means “[someone thinks that] you [in fact] do X,” whereas the normal usage means “You should do X [which you are not in fact doing].” And (in my dialect, anyway) there’s a phonetic difference as well: the passive is /səˈpoʊzd/, whereas the normal usage is /səˈpoʊst/.

  130. David Eddyshaw says

    Yes, there seems to be a regular cottage industry of English past participles deciding that they want to break out on their own and become Real Adjectives. They may decide to change their (phonological or syntactic) look to symbolise their new lives.

  131. @DE, I hope we both agree that the adjectves … participles is a scale or a range. On the left we have adjectives that feel and behave as adjectives, but have the form of a participle. Often a form is somewhere between the extremes. It retains some verbal properties.

  132. David Eddyshaw says

    Indubitably.

  133. David Eddyshaw says

    I see that CGEL does in fact discuss this very issue of stative versus dynamic in passives (p1437f.) They point out that what they call “adjectival passives” are indeed always stative, but go on to warn that

    It must be emphasised, however, that adjectival and verbal passives cannot be distinguished simply by asking whether the interpretation is stative or dynamic – it is for this reason that we have not included it among the tests for adjectival status.

    They make two points: firstly, verbal passives may have a stative meaning if the verb itself is stative: “She is loved by everyone.” Secondly, adjectival passives can function as predicative complements of dynamic verbs: “it became magnetised.” (Neither of these cases apply with the “sat”, “stood” or “lain” types we’ve been talking about, though, so for the purposes of our discussion it probably is a valid test.)

    I hadn’t really thought about this issue of aspect in adjectives in this context much before, although in retrospect my not-very-relevant remarks about aspect in Kusaal deverbal adjectives above were my subconscious trying to bring it to my attention. Kusaal in fact lacks precisely the kind of deverbal adjective that corresponds to our participles: this is obvious with present participles, but is actually also true with forms like kɔɔlʋŋ “broken”, from the intransitive verb “break”; this does indeed mean “broken”, but only in the adjectival sense of the two possibilities mentioned in CGEL’s section on “Verbal passives and adjectival passives” that I cited above. (Not for the first time, free-associating about these issues on LH and the helpful input of fellow-Hatters have helped me understand Kusaal better.)

    I think that one can say of English core adjectives (i.e. not pariticiples), that like all Kusaal adjectives, they can only be aspectually habitual or stative, not progressive or perfective. Part of the journey for a participle of an English dynamic verb setting out to become a “real” adjective is shedding any other aspectual quality than habitual or stative.

  134. David Eddyshaw says

    My mistake of putting “lain” for “laid” in the above comment is my subconscious being helpful again …

    “I am laid on a beach towel.”
    (equivalent to Soanesian “I am lying on a beach towel”, though it could be timeless/habitual passive: “Who lays you on a beach towel? Why?? What nameless rite is this?”)

    “I am lain on a beach towel.”
    Wardour Street English for “I have lain on a beach towel.” “Fie! how improper of you, Jane. You Jezebel!”

    So: in this case, the “sat” construction requires the past participle form of the transitive-causative “lay”, and doesn’t work with the past participle of the intransitive “lie.” So it must have originated via a shift from verbal/dynamic passive to adjectival/stative passive, and cannot possibly be a survival of an intransitive be-perfect.

    Therefore, if this is indeed analogous to “I’m sat at a table”, “I’m stood at the bar”, which seems very probable to me, the unSoanesian “sat” construction did indeed develop from a passive (and might still be called “passive” at a pinch, though as CGEL says, this is a bit of a stretch for the terminology), and has nothing to do with the be-perfect and the construction seen in the Anglo-Saxon Chronicle.

    [I have studiously avoided citing any “get” passives of “lay” here, as ambiguity often offends and this is a family blog.]

  135. @DE, what I just wrote was the first paragraph of a very long comment meant to explain why I can’t agree with your “makes no sense whatsoever to call a construction “passive” if there is no corresponding active construction”.

    I think I must edit (and shorten) it before sending, but in light of what you just wrote, the second paragraph:

    ‘some examples:

    рассерженный “angered”. In this case there is a corresponding verb рассердить “to make angry”.

    When I use “рассерженный”, I still can feel its perfective aspect (from the prefix): the person underwent a transition from calm to angry. I don’t feel that the described person is a patient of something. This may have to do with both existance of intransitive рассердиться “to get angry” and the causative semantics of сердить (and general understanding that the person who angered you might have no intent to do that) – but I do feel that there was an external trigger.

    “Aspect” is a verbal property. Thus when I use it, I can use it attributively as I use adjectives, but it still retains some verbal semantical properties. ‘

  136. David Eddyshaw says

    Russian may very well work quite differently from English in this regard, as in so many others. I don’t begin to know enough about Russian to say anything about it, really.

    I will say that the Russian aspect system (despite being the poster child for grammatical aspect) actually works rather differently from English aspect, which is much more like the systems you typically find in West Africa (and in Biblical Hebrew, too.) In particular, in West African systems the perfective is normally the default unmarked aspect, not the imperfective as in Russian. English uses the same forms for perfective and habitual (the “simple present”, in the present tense) but otherwise has a system much more like Kusaal than Russian.

    Aspect surely is primarily a verbal property; nevertheless, I think it does make sense to talk about aspect in adjectives, at least in some languages (especially, of course, those like Classical Greek with exuberant systems of participles.)

    I think you can reasonably ascribe aspect to deverbal nouns (like gerunds) in some languages, too.

    [I see have gratuitously muddied the waters by using “stative” in two different senses above, too. They overlap but are not the same thing. English “stative verbs” in their usual senses can’t form continuous/progressive tenses; but “stative” as a property of constructions is an aspect and even dynamic verbs can be used in stative constructions. Apologies …]

  137. @DE, what I think is that “passivity” (that is being a patient) is a semantical property, moreover, a “verbal” semantical property. It can be felt (not in the word рассерженный).

    What if a word is “between” the two extremes (which may mean: (1) is not fully independent of verbs, (2) has certain verbal properties) but there is no corresponding active construction?

    It seems to exclude (1) “is not fully independent” but it does not seem to exclude (2). At least a claim “when there is no active construction, passive semantics must be absent from the adjective/participle” does not sound as something a priori true. It sounds like a generalisation that can be dispoven or confirmed by observations.

  138. “Russian may very well work quite differently from English in this regard”

    DE, true. But “makes no sense whatsoever to call a construction “passive” if there is no corresponding active construction” sounds as a general terminological proposal (a definition of “passive”) applicable to any langauge…

    I am just explaining why I still may need the word “passive” in absence of verbs. “Patient” is a semantical property. “Passive” participles tell that the referent is a patient of some action.

    Also paradigms have gaps. It is normal…

  139. If we need examples:

    мы помолвлены “we’re engaged”
    мы обручены “we’re engaged (with specifical reference to the exchange of rings)

    We also have reflexive мы обручились (we [exchanged-rings-[as-a-symbol-of-engagement]]) and there is an uncommon transitive обручить, but I don’t say “нас обручил Х” because there was no X:-)

    No one around me says “помолвились” or “помолвил”. Still I reconstruct this verb with such an ease that I am sure that some speakers must use it (we can google помолвились and see if I’m right).

    And I think the ideas of both “change that happened in past” and “something was done TO us” are at some level present in помолвлены.


    In English I would suspect some verbal properties in “destined to…” and “born to…”, but the semantical side of it (if it exists) only accessible to native speakers (otherwise formal combinability with “to” is a verbal property, of course)

  140. David Eddyshaw says

    I’m distrustful of properties that can only be “felt.” Sprachgefühl, like its cousin Common Sense, is not always a reliable guide.

    I also think it’s important (do as I say, not as I do!) not to get confused between form and meaning.

    In this particular case we’re talking about forms which – as they stand – are indistinguishable from passives. The feeling one has as a native speaker that they are not actually passive, but a distinct construction, is an important starting point for investigation (especially when, as here, actual significant ambiguity is detectable), but unless you can correlate this subjective difference with potential formal differences you haven’t really got anywhere with the analysis yet.

    In this particular case, there turn out to be quite a number of formal differences once you start playing with the forms a bit: if you interpret these “passives” in one way, you find that you can’t inflect them for progressive aspect without losing that possible interpretation, for example. The possibility of adding an agent or its impossibility is just another way of playing with the construction to find out if your subjective impression (that there are two distinct constructions involved which just happen to look identical in some cases) is really objectively valid. In this case, this too obligingly turns out to correlate with the subjective meaning difference quite neatly.

    None of this means that everything can be fitted into neat boxes. Of course it can’t: as the Wise Grammarian said: All grammars leak.* But that doesn’t mean that we can’t achieve anything at all.

    At the end of the day it doesn’t really matter whether you call a construction “passive” or not, so long as you’ve explained clearly enough what you mean by “passive” for the purposes of your investigation and displayed your evidence properly (which is exactly what CGEL does in the passage I quoted.)
    .
    Though it’s a kindness and courtesy to your readers not to use terms in such a gratuitously nonstandard way that they can’t understand you without looking up your idiosyncratic definitions at every turn … **

    * In this particular case, the leakage is manifest in Latin and Greek grammars with their “deponent” verbs. They’re passive … except they’re not … but if we slap a name on it, we can plug that leak. Maybe.

    ** Hands up everyone who immediately understood why CGEL says of the “verbal passive” The vase is broken: “broken is a bare passive clause consisting of just the head, the verb broken“? CGEL is actually very consistent in its analyses of such things, and does explain what it’s doing very clearly, but I do feel there is a certain Hermetic quality to some of its terminology at times …

  141. David Eddyshaw says

    otherwise formal combinability with “to” is a verbal property, of course

    Not at all; “near to”, “superior to” …
    If you want “to” preceding an “infinitive”: “prone to”, “liable to” …

  142. that they can’t understand you without looking up your idiosyncratic definitions at every turn …

    and

    In this particular case, the leakage is manifest in Latin and Greek grammars with their “deponent” verbs. They’re passive … except they’re not … but if we slap a name on it, we can plug that leak. Maybe.

    @DE, linguistics does recognise existance of semantics. I know that some people believe that syntax IS linguistics – but even Chomsky recognise that semantics exists. Thus your “deponent” verbs.

    My definition is entirely normal.

    WP:

    “In a clause with passive voice, the grammatical subject expresses the theme or patient of the main verb – that is, the person or thing that undergoes the action or has its state changed.[2] This contrasts with active voice, in which the subject has the agent role.”

    Emphasis mine: agent and patient are semantical roles.

    WP:
    “Some languages (such as Latin and Russian) have distinct participles for active and passive uses. In English, the present participle is essentially an active participle, while the past participle has both active and passive uses.”

    What do you think they mean by “passive use”?

  143. PlasticPaddy says

    @dravi
    Maybe they mean
    I have been hurt (passive)
    Ihave hurt myself (active)

  144. Lars Mathiesen (he/him/his) says

    Those guys from Alpha haven’t really figured out how to make humans at ease yet.

  145. David Eddyshaw says

    @drasvi:

    I wasn’t aiming the “idiosyncratic definition” thing at you (CGEL, wonderful though it is, is a much more suitable target for that.)

    If anything, I think that your definitions aren’t idiosyncratic* enough

    I don’t think “active” and “passive” are labels that you can simply apply across all languages in the same sense. “Agent” and “patient” (even) are quite slippery notions once you start comparing across languages, for all that there are obvious huge areas of common overlap, and as for “subject” and “object” …

    Is the subject of “love” an “agent”?

    Kusaal can make an “agent” noun from the copula … (“one who is something …”)

    * You say that like it’s a bad thing …

  146. linguistics does recognise existance of semantics

    Of course, but you seem to be saying semantics is the only thing that matters.

  147. David Eddyshaw says

    Illustration of the difficulty with “active” and “passive” from the One True Language (other than Welsh, obvs, which actually also could provide examples of Problems with the Passive):

    M nu daam la. “I’ve drunk the beer.”
    M nuud daam la. “I drink the beer.”
    M nuud nɛ daam la. “I’m drinking the beer.” (The particle after an imperfective makes it continuous/progressive.)

    Daam la nuya. “The beer has been drunk.” [This -ya is automatically attached to clause-final perfectives in positive indicative main clauses: it’s not an aspect flexion or a voice marker but part of the system for marking (in)subordination.]
    Daam la nuud. “The beer gets drunk.”
    Daam la nuud nɛ. “The beer is for drinking.” (Not, e.g. washing the car with. The sentence cannot mean “The beer is being drunk”, and the particle cannot be interpreted as marking continuous/progressive aspect but is forced into the role of constituent focus particle.)

    Are the latter three sentences “passive”? There is no way to express an agent, but that’s true of passives in quite a few languages, in fact, so that doesn’t rule out a passive interpretation. There’s no formal mark of a voice change, but I suppose that’s not conclusive either. But why can’t the “passive” forms have continuous aspect? Contrast

    M bɔdig daam la. “I’ve lost the beer.”
    M bɔdigid daam la. “I lose the beer.”
    M bɔdigid nɛ daam la. “I’m losing the beer.”

    Daam la bɔdigya. “The beer’s got lost.”
    Daam la bɔdigid. “The beer gets lost.”
    Daam la bɔdigid nɛ. “The beer is getting lost.” (No problem with as continuous.)

    Is this “passive”? If so, why does it behave differently aspectually from the first set of examples?

    Lots of West African languages do this sort of thing in fact, though the particular aspectual restrictions are a Kusaal peculiarity; it may be more widespread in Oti-Volta, but the existing grammars don’t go into enough detail to tell.

    Swahili (like many other Bantu languages) does it, not with objects, but with locatives
    Would you like to hear about Tagalog?

  148. Do these examples work as well with the verb for ‘buy’ (which to my mind is even more clearly instantaneous than ‘lose’)?

  149. Well, sorry if I sounded angered in my comment above. I do not feel so (though one may say that what mattes is my distribution and not my feelings…):(

    I did think that “idiosyncratic” refers to my understanding of passivity. But it is not why I felt the need to protest. DE correctly corrected my use of “impersonal” above and he was right!

    I protest against the idea that Serious linguistics is Syntax. I misinterpreted DE as saying that “passive” is a primarily syntactical concept intended for “passive constructions”.

  150. @LH, conversely.

    I explained why I need this word for describing semantical properties.
    But I am not saying that it is an error to speak about “passive constructions” and otherwise apply it to syntax.

    For all I know, the concept is rooted in both syntax, semantics and morphology.

  151. David Eddyshaw says

    Do these examples work as well with the verb for ‘buy’

    Da’ “buy” behaves like nu “drink.”

    I protest against the idea that Serious linguistics is Syntax

    Me too (though I do find syntax very interesting.) I think we’re fundamentally in agreement …
    In particular, I think that the attempt to understand syntax while sidelining meaning is the Original Sin of Chomskyism. I think form and meaning are firmly linked all the way down, from text to morpheme: so you can’t meaningfully (hah!) study one without the other.

  152. David Marjanović says

    Yup, I had to learn the English and the Russian aspect system completely separately, from scratch; they’re orthogonal to each other.

    The Slavic composite past with “be” (or without it in Russian, where “be” has basically disappeared, or in the 3rd person in Polish…) is a great mystery to me at least in historical terms, because the participle it uses (in *-l-) doesn’t occur as a participle anywhere else in IE, and there’s not even any room for it to occur outside that tense in the Slavic system that already has four other participles (present, past × active, passive). I don’t think this has much to do with the Romance and Germanic composite perfects that ultimately seem to come from a Greek construction with an active and/or aorist participle (I forgot, and I don’t know enough Greek to reproduce the idea).

    At least some of the “South Slavic” languages also retain simple past tenses; Noetica’s Bulgarian example is an aorist AFAIK.

    “when I’m gone” is an interesting one

    That’s “gone” fossilized as an adjective meaning “away”, I would say.

    “Christ is risen” is a formula borrowed from Early Modern English

    Twice I wanted to say that, and both times I decided against it because I’m wondering if it’s been reinterpreted as an adjective of that sort…

    Tathāgata

    The galaxy-brain pun.

    Also the “proper” causative of stå seems to have evaporated; I don’t have the reconstructive nous to posit a possible PG form and carry it forward.

    Actually, it’s stellen in German, so the extreme irregularity of it all can be safely dropped into the memory hole of time immemorial. Though Stall “stable” has to be related somehow. It never occurred to me that still “silent” should be related, too – but it should…

    (We also have the jocular “was disappeared”-type forms ist gegangen worden and ist zurückgetreten worden [“was resigned”], but that must be a lot more recent…)

    I’ve often wondered if a shorthand system based on sound such as Pitmans could be adapted to become a writing system for all languages.

    Of course, for some value of “adapted”.

    Canadian Aboriginal Syllabics is in fact developed from Pitman’s.

    This question of how derived adjectives relate to cognate verbs aspectually is quite interesting in its own right, I think.

    Oh, that reminds me. My dialect has lost the present participle – except for a handful of lexicalized relicts of which some have become ordinary adjectives but others are used in a non-productive mini-aspect that doesn’t exist in Standard German: *wenn es kochend/regnend wird “if/when it starts to boil/rain”. “Raining” doesn’t seem to be used anywhere else in fact.

    And (in my dialect, anyway) there’s a phonetic difference as well: the passive is /səˈpoʊzd/, whereas the normal usage is /səˈpoʊst/.

    And that’s not neutralized by the following to?

    (…If not, that might actually explain that commenter on a webcomic out there who always spells “have to” haft to. Partly anyway.)

  153. I first saw “was disappeared” in Catch-22, a repeating ominous occurrence.

  154. And that’s not neutralized by the following to?

    Nope; if it’s the (rare) genuine passive, I would pause momentarily before the “to” to prevent that assimilation and make the sense clear. But I’m not sure I’ve ever actually used that in speech. It’s quite rare.

  155. David Eddyshaw says

    I’ve seen hafta for “have to” quite often as eye-dialect representing Low-Status White American (in texts written by Not-so-low-Status White Americans.)

    I don’t think I devoice the /v/ in this sequence myself, but introspection on such matters is hardly a safe guide …

  156. DM:

    At least some of the “South Slavic” languages also retain simple past tenses; Noetica’s Bulgarian example is an aorist AFAIK.

    You’re right of course. I should have chosen a different example like Polish, Russian, or Czech. The indicator would be variation by gender (as in southern forms that retain the auxiliary). Czech:

    Nakreslil kruh.
    He drew a circle.

    Nakreslila kruh.
    She drew a circle.

  157. @David Marjanović: And that’s not neutralized by the following to?

    No, and this is often the case when dealing with two words that are transparently etymologically identical, but not quite homophonous.* Since the adjectival construction with supposed to is so much more common, using the true passive version invites particularly clear enunciation. If the voicing is completely neutralized, it just sounds wrong. The /z/ has to remain voiced, although the /d/ can potentially be devoiced to /t/. On the other hand, if a speaker’s enunciation is especially strong, the /d/ could feature anything up to full fortition.**

    * Have I mentioned that homophone is my least favorite sounding English word? I don’t tend to develop word aversions—and there are no words that I dislike the “sound” of because of what they mean. However, from the day that I first encountered homophone in third grade, I have disliked the way it sounds. My dislike has diminished quite a bit over the intervening decades, but it is still there a bit. Conversely, my favorite-sounding English word is erythropoietin.

    ** In English, because there are no normally voiced aspirated stops, and lenis-fortis distinctions are basically never phonemic, a /d/ can be emphasized*** (especially at the end of a word or syllable) by fortition to /dʰ/. The same goes for the other stops, although I think it is less common with them. But is this true in all Germanic languages?

    *** “Fortisfied”?

  158. David Eddyshaw says

    Fortitionerated.

  159. David Eddyshaw says

    This might have contributed to my apparently-mistaken perception of them [Noetica] as female. Though the last letter of the username was probably more important.

    Neuter plural, obviously. (Same as Us.)

    [On the Internet, nobody knows you’re an eldritch abomination.]

  160. AntC on remained sat/stood: “Those examples I’d call acceptable but a bit jarring: ‘remain’ is formal register; ‘stayed sat’ sounds more natural.”

    Excellent point, my American brain was blind to that nuance. It’s OK for CGEL since they’re only trying to cover fairly standard forms of the language, but I can see how it could interfere here. Here are some examples of “stayed sat/stood”. Do they sound natural?

    What’s the longest you’ve stayed sat in an uncomfortable position to avoid waking a partner who has fallen asleep on you? (twitter; Warwick, England)
    it was a Christmas present but with nowhere to go has stayed sat in it’s box until today! (Facebook; Hampshire, England)
    Some stayed sat, James stayed sat, he didn’t know what else to do, what the rules were, what would happen if he got up (self-published fiction; British)
    We looked for our pulse either on our neck or on our wrist. We stayed sat down and counted how many beats we could feel. (children’s school website; Lancashire)
    I was sat alone in my bedroom. I felt claustrophobic, as if the space which I could potentially occupy in the limited time I had left was shrinking … I stayed sat that way for a long time. (blog; Manchester)
    Knight’s music is a mix – much of it is sit-down music, whilst some of it is groovy, yet most of the audience stayed sat for the full duration of the show. (The Mancunion)

    they then stayed stood about until paramedics arrived which was around 20mins. (TripAdvisor review; Middlesbrough, UK)
    Ruud came a bit later and Djokovic stayed stood talking to them for ages after finishing his food before he left the site for the day (twitter; sportswriter, Brighton, UK)
    Molly’s Lips followed Katie and I [a band], but everyone stayed stood for them. (music review, Liverpool)
    They had gone to wrong address even though clarified where we were, i stayed stood outside in cold (just-eat.co.uk)
    I stayed stood there until hubby was almost out of sight. (blogger, Manchester)
    Mrs Ridgard replied: “Yes, we all got down, apart from my boss Jeffery (Walton) who stayed stood up.” (Yorkshire Evening Post)

    There are other verbs that take predicative complements, chiefly look, appear, seem; I would guess these won’t work with sat/stood? (Of course, there’s stuff like “that sandwich looks sat on”, but that’s different.)

  161. What’s the longest you’ve stayed sat in an uncomfortable position to avoid waking a partner who has fallen asleep on you?

    I was asking myself that only yesterday. All those examples seem entirely natural.

    Hartlepool: come for the shipyards, stay sprawled for the beaches.

  162. They had gone to wrong address even though clarified where we were, i stayed stood outside in cold

    What’s going on with that “clarified”? Is it a passive, with a nonstandard object; that is, would an example of a corresponding active construction have the object of the transitive clarify be the person to whom the explanatory information is being given?

    Of course, clarify can be transitive, with roughly the same semantics as make clear:

    The next thing the recipe says is we should clarify the butter.
    I paused to clarify the definition for the class again.

    but not

    *I wrote out the definition again, to clarify the class.

  163. @ DE: [I have studiously avoided citing any “get” passives of “lay” here, as ambiguity often offends and this is a family blog.]

    Gee, and I was very much looking forward to that analysis with the beach towel example.

  164. Maybe, or it’s a typing error for “even though I clarified”. It’s a pizza review that looks like it was tapped out on a phone in a hurry, with some words dropped.

  165. I often hear constructions such as “I’m sat here” where the meaning is to establish exactly where I’m sitting, not where somebody has seated me. It’s not passive. It’s simply idiomatic use of the past tense and not standard English.

  166. The vase was broken. [be-passive or complex-intransitive]

    As a passive (of the short variety) this describes an event, as does the active clause Someone broke the vase. As a complex-intransitive clause it describes a state – the state resulting from someone or something breaking the vase. In the first interpretation broken is a bare passive clause consisting of just the head, the verb broken; in the second, broken is an adjective.

    “The show is over” is suspiciosly similar…

  167. Trond Engen says

    David E.: Kusaal makes inchoatives out of body-position verbs (e.g. “sit down” from “sit/be sitting”) with a derivational suffix -n

    Needless to say, so does Scandinavian. Not on body-position verbs, synchronically, but arguably borderline productive on root adjectives:.

    Established pattern:
    lys/i> a. “light”, lysne v. “become light”
    gul a. “yellow”, gulne v. “become yellow”
    syk a. “ill”, sykne v. “become ill”
    myk a. “soft”, mykne v, “become soft”
    hard a. “hard”, hardne v. “become hard”
    etc.

    (Eng. lighten, harden, soften, etc. are regularly transitive. That’s not automatic in Scand.)

    Applied to new adjectives:
    røff a. “rough” (> Eng “rough”), ?røfne v. “become rough”
    kul a. “cool” (> Eng “cool”), ?kulne v. ” become cool”
    sløy a. “slick” (> Eng, “sly”), ?sløyne v. “become slick”

    which not only appears nowhere else in derivation

    Also on some noun roots:
    dagne v. (arch.). “dawn”

    Historically it has to be (at least partially) patterned on verbs derived from the perfect participle in -n(-).

    drukne v. “drown”
    tagna v. (Nyn.) “become silent”
    segne v. “fall together, faint”

    but is the only derivational suffix which regularly alters the tones of the root it’s attached to.

    No, that’s confined to the Trans-Saharan branch.

  168. Jen in Edinburgh:

    Is this the kind we’re looking for?

    1617 Misc. Abbotsf. C. 311.
    This witt was sett smoaking … in ane obscure chymney corner at Leethe

    OED also has that under sit, in the form history for one of the Scottish branches of the past participle, but they don’t use it for any of the definitions, so we don’t know how they interpret it. It could theoretically be a be-perfect, or a true passive of set “without causal implication.” The context is from an “Answer to the Satire Against Scotland” attacking some satirist:

    What pitie it is, that this witt was sett smoaking among the dryed whyttings in ane obscure chymney corner at Leethe: he could haue taught ʒow ane hundrethe fyne knakis to entertayne ane king, schowis and pageantrie, and puppett playis : if he would haue bein but a lytle idolatrous, he had mad ʒou famous for ewer. As it was, he confessed a great schowe off wynen of charge ; ʒour weaponschawe and ʒour fireworkis cost ʒou nothing, whilles (I warrant) this snake will newer be worthe the wearie coales ʒou spent in bonfires.

    … well, I don’t really follow that, but I don’t think this can be equivalent to modern standard English “was sitting smoking”; it sounds like it’s most likely a passive of set.

  169. Thank you @ktschwarz for those excellent examples. Does your repository include Aus/NZ data?

    Away up-thread there were suggestions stative ‘I’m sat/stood’ also appeared there. I noticed at one workplace I visited in NZ there was someone who almost went out of their way to use it. I put it down to him having UK parents, although there was nothing Brit about his accent, and his workmates didn’t bat an eyelid.

  170. David Eddyshaw says

    Needless to say, so does Scandinavian

    Thanks, Trond.

    In view of such detailed correspondences at every level of the grammar, it is remarkable that Ethnologue continues to misclassify the Scandinavian branch of the family. There can be little doubt that the so-called “Germanic” features of these languages are due to contact.

  171. PlasticPaddy says

    @trond
    the corresponding transitive verbs in German would seem to be formed as be-[NOUN STEM]-ig-en, where en is the standard infinitive suffix. Some of the [NOUN STEM]-ig forms are adjectives, but not., e.g., *friedig, *sänftig. There are also some forms with er- instead of be-. Although English forms like harden, soften, moisten seem to be formed directly from the adjective, I think it is possible there were originally noun forms analogous to German Härte, Sänfte etc., that formed the basis for the English transitive forms (compare AS (be)notian = “make use of, employ, enjoy”, (be)sidian = “make or to become wide, ample”).

  172. Trond Engen says

    Yes, the infinitive suffix is surely in the mix, as is the -n- of the present participle. The duality transitive/inchoative (esp. in English) could well be due to a lost suffix.

  173. January First-of-May says

    Neuter plural, obviously. (Same as Us.)

    [On the Internet, nobody knows you’re an eldritch abomination.]

    Unless they tell you; I’ve met a few that did, though of course maybe some of them weren’t and just wanted me to think they were.

    However, as it happens, my linguistic competence (such as it is) is rather limited on neuter plurals. English and (post-1917) Russian don’t (appear to) have any gender agreement on plurals, and AFAIK the former has no neuter as well; French and Hebrew have no neuter, and use the masculine plural as the neutral form.
    Unfortunately my meager knowledge of German is not enough to tell what (if anything) it does with neuter plurals, and my even more meager knowledge of Spanish is not enough to tell whether it has any neuter forms. Latin probably has some neuter plurals but offhand I can’t think of any…

    (Kusaal has no gender at all – or, rather, it has about a dozen, which bear no relation to human genders. Welsh is in the same category as German.
    Pre-1917 Russian distinguished masculine and feminine plurals, but still not neuter, AFAIK.)

     
    EDIT: wait, is being an abomination a necessary condition of being eldritch? The aforementioned beings I have previously conversed with online certainly seemed quite eldritch, but I wouldn’t have called any of them an abomination, as such, so I’m not sure if the term quite applies.

  174. Trond Engen says

    Speaaking of the present participle, was sat etc. reminds me of two different-but-related Norw, constructions:

    1. ble + pr.part.:

    Jeg var så trøtt at jeg ble sittende i bilen mens de andre gikk og handla.
    “I was so tired that I stayed in the car while the others went shopping”

    Vi kjøpte plankene, men de ble liggende i kjelleren i årevis før vi satte i gang.
    “We bought the planks, but they were left in the basement for years before we started the job.”

    De kom en time før det begynte, men likevel ble de stående i kø halve formiddagen.
    “They came an hour before it started, but still spent half morning waiting in line.”

    Ingenting var klart da vi kom, så vi ble gående og slenge i flere dager.
    “Nothing was ready when we came, so we were just hanging idly around for days”

    2. kom + pr.part.:
    (This translates more directly to English)

    Han kom kjørende med hele familien i baksetet.
    “He came with his whole family in the back seat of the car.”

    De kom flyttende med en gang.
    “They moved here immediately.”

    Vi sto der og ante fred og ingen fare, og så kom plutselig læreren gående.
    “We stood there suspecting nothing, and suddenly the teacher showed up.”

    I don’t have time to sort out the aspectual difficulties.

  175. January First-of-May says

    (post-1917) <…> Pre-1917

    1918, of course. (The reform was introduced in 1917, by the Temporary Government, but not implemented until 1918.)

  176. David Eddyshaw says

    Kusaal has no gender at all – or, rather, it has about a dozen, which bear no relation to human genders. Welsh is in the same category as German

    As far as agreement (which appears only with pronouns) goes, Kusaal has just animate versus inanimate genders, and the plural makes no gender distinctions at all. The system is “natural”, meaning that you can simply infer the agreement gender from what the noun refers to in the real world. The inherited noun class system is alive and well, with five count-noun classes and two mass-noun, but only as morphology: there is no actual agreement by noun class any more, though there are a few stranded relics of it here and there. As in Niger-Congo everywhere, actual sex has never had any consequences for how the system operates; one class is “human”, in the sense that all nouns in that class refer to people, but it doesn’t include all human-reference nouns. Elsewhere in Niger-Congo that class has often expanded to include all sorts of non-human nouns, quite often as the default class that loanwords go into, but no Oti-Volta language has done that so far.

    Loss of agreement is an areal thing within Western Oti-Volta; agreement of pronouns, adjectives and numerals by noun class is very much still the thing elsewhere else in Oti-Volta, and in the geographically separated Western Oti-Volta language Boulba (way over in Benin) it’s still fully operational. Proto-Western must have had class-based agreement gender, and it’s been lost in the individual languages after they became distinct from one another.

    Welsh has masculine and feminine as grammatical genders, but makes no gender distinctions in the plural. You have to go back to Old Irish for a neuter in Celtic.

    There are quite a few Latin neuter plurals in English: data, media, Americana, arcana …

    is being an abomination a necessary condition of being eldritch?

    Well, these things depend rather on your point of view. And on the whole, I prefer to appear in a form people are comfortable with.

  177. @January First-of-May: “The Seven Geases” demonstrates that an entity may be eldritch while not being an abomination, an abomination but not eldritch, both, or neither.

  178. “[On the Internet, nobody knows you’re an eldritch abomination.]”

    I would not be so sure. (substituting me for “you”).

  179. Re neuter plurals in German: In German, no gender distinctions are made in the plural of adjectives and pronouns. Of course, neuter nouns have plural forms, and there is a plural formation that is more typical for neuter nouns (umlaut plus -er), but it’s also used for male Mann / Männer, and there are many neuter nouns that have other plurals.

  180. January First-of-May says

    Re neuter plurals in German: In German, no gender distinctions are made in the plural of adjectives and pronouns.

    So it’s like (modern) Russian, then. About what I expected (vaguely recall that this might be a SAE thing?) but I wasn’t sure.

  181. David Marjanović says

    But is this true in all Germanic languages?

    Some – in particular almost all of southern German – lack voiced obstruents entirely, so I’m still unfamiliar enough with the whole concept that I have to ask about the last few details.

    [b d g] are the most difficult sounds in the entire French language, followed by [z ʒ]…

    dagne v. (arch.). “dawn”

    Oh. German (arch.) actually has tagen “dawn”, but it never occurred to me it’s the cognate of dawn!

    (There’s another tagen, which remains in common use, but it means “be in session”.)

    Generally in German, this -n- seems to have been completely confused with the infinitive/1pl/3pl ending, except for rechnen “calculate, reckon” < noun prefix Rechen-* and zeichnen “draw, sign” < Zeichen “sign”.

    *…but not Rechen “rake”.

    segne v. “fall together, faint”

    Likewise it never occurred to me that sacken (zusammensacken) is the cognate of sag, despite the identical meaning. The ancient loan Sack must have gotten completely in the way.

    and my even more meager knowledge of Spanish is not enough to tell whether it has any neuter forms.

    Nope. Basic Romance – everything is m or f, and that’s kept in the pl.

  182. David Eddyshaw says

    About what I expected (vaguely recall that this might be a SAE thing?) but I wasn’t sure

    No, I think it’s a cross-linguistic tendency. Fulfulde has twenty singular noun agreement classes and “only” five plural.

    Hausa distinguishes masculine and feminine grammatical genders in the singular but not in the plural, and this seems to go all the way back to Proto-Chadic (a lot of Chadic languages have given up grammatical gender altogether, though, including the whole of the Biu-Mandara/Central branch.)

    Dagaare, which like most of Western Oti-Volta has abandoned noun-class-based grammatical gender for a natural animate/inanimate system, bizarrely only distinguishes gender in the plural. I can’t actually think of any other language which distinguishes more agreement genders or classes in the plural than the singular. Weird.

  183. It is interesting that plurals almost always show the same or fewer gender distinctions than singulars, when ‘logically’ there would be an obvious motivation for having more, to cover groups of mixed gender (particularly if you distinguish different human genders in the singular). Dropping distinctions in the plural allows you to sidestep the issue, but in some ways it is odd that no one traditionally has a contrast of m/f in the singular versus m/f/mixed in the plural.

  184. David M., David E.: Neuter nouns are indeed unknown in Welsh or Spanish, but Modern Spanish does have neuter forms directly inherited from Latin: neuter singular forms of the definite article and of the demonstratives (when used pronominally): they are distinct from masculine and feminine ones, and their forms go back to Latin: “lo” for the neuter article, “el” for the masculine and “la” for the feminine article (all singular), for example.

    David M.: Actually, Modern Romanian has a full-fledged neuter nominal gender (AKA “ambigeneric”), which is alive and kicking, as is shown by the fact that numerous modern anglicisms are assigned this gender. The neuter also remains alive in a number of other Romance varieties (not in the national languages other than Romanian, admittedly).

    (However, not all that glitters is gold: (some forms of) Asturian could be said to have a neuter (AKA “mass gender”), but it has recently been shown, conclusively to my mind, that this must be an Asturian innovation, due INTER ALIA to analogical extension to adjectives of the neuter /o/ ending which must originally have been found solely on the definite article and demonstrative pronouns, just like in Spanish).

    It is thus clear that the neuter (as a noun gender) was alive and well in Proto-Romance, and that its disappearance postdates the loss of Romance (or even Italo-Western Romance) linguistic unity. Incidentally, neuter forms can also be found in adjectives in some modern or attested Romance varieties: In Old French there was an opposition between nominative singular masculine CLERS, feminine CLERE, and neuter CLER, for instance (from CLARUS/CLARA/CLARUM, respectively).

    David E.: In this light, Middle Welsh had neuter forms of demonstrative pronouns (identical to plural forms, but in Proto-Brythonic masculine singular, feminine singular, neuter singular and plural forms were each distinct from the other three), so technically you do not need to go all the way to Old Irish to find neuter FORMS, as opposed to neuter nouns in Celtic.

    (Incidentally, the loss of the neuter in Brythonic must have taken place very shortly before its break-up: the oldest stratum of Latin nouns in Old Irish must have been borrowed indirectly, via Brythonic (the phonological evidence is unequivocal) , but the Latin gender is (as a rule) preserved in Old Irish, including the neuter: for the Latin neuter nouns to have become neuters in Old Irish despite being borrowed by way of Brythonic only makes sense if you assume Brythonic, at the time it mediated these loanwords, still had a nominal neuter).

    Whether the typological similarity between Proto-Brythonic and conservative Romance languages such as Spanish (both with a binary masculine/feminine nominal gender system, and both with a neuter form found with pronominal demonstratives which also had distinct masculine and feminine forms) is a coincidence or a sign that Proto-Brythonic was more heavily Romance-influenced than has perhaps been hitherto acknowledged remains unclear. I freely admit I lean towards the second alternative.

  185. David Marjanović says

    At least some of the older Germanic languages used the neuter plural for mixed groups… because, by coincidence, the masculine dual had ended up as identical with the neuter plural before it otherwise disappeared from declension (before Proto-Germanic times).

    Modern Romanian

    That phenomenon (also found, unproductive IIRC, in Italian) is what I tried to talk around by saying “basic”… :-] Yes, I’ve seen the Spanish generic pronominalized article in action (lo que es “that which is”).

    The neuter also remains alive in a number of other Romance varieties

    Are they all “ambigeneric”, or…

    In Old French there was an opposition between nominative singular masculine CLERS, feminine CLERE, and neuter CLER, for instance

    Of that I had no idea. That’s fascinating.

  186. David M: The English cognate of German tagen is not dawn, but daw (now obsolete except in Scotland). How that turned into dawn is slightly complicated: daw (v.) > dawing (n.) > somehow dawning (n.) > dawn (v.) by back-formation.

  187. David Eddyshaw says

    Jarawara (the Arawa language) has masculine and feminine grammatical genders, gender assignment being essentially unpredictable for inanimates; it uses feminine plural for mixed groups. Feminine is the default gender for words like “child” if the sex is not known, and all pronouns, regardless of their actual reference, take feminine agreement.

    (Also, inalienably possessed nouns take the gender of their possessor; one of those things where, after being initially nonplussed, you eventually go “Yeah, I see that.”)

  188. David Marjanović says

    > somehow dawning

    Oh, following the link I see the Vikings can be blamed. Excellent.

    Feminine is the default gender for words like “child” if the sex is not known

    Compare Geschwister “sibling(s)” and a similar phenomenon in Finnish that I shouldn’t look up at this hour.

  189. Noetica ők, őket, övék says

    And girl originally meant a young of your species of either gender, as has no doubt been noted innumerable times before chez Chapeau.

  190. David Eddyshaw says

    Is Hungarian a Form they are Comfortable With?
    (Asking for some friends.)

  191. @DE, I checked CGEL, and I don’t understand some things.

    (1) what are adjectives and participles (or in their syntax-freindly formulation “verbal passives” and “adjectival passives”)? They offer a test (‘very’, substituting ‘be’ with ‘become, look, seem’) but these are tests. They help you classify something, but they don’t explain where this distinction comes from.

    Is “reporting an event” vs. “state” the base of distinction or what? (it is actually a somewhat subtle distinction, I already mentioned Russian words like устал “[am/got] tired” and other words where -l is productive participial suffix, where I often can’t tell if I am reporting an event or state as a native speaker!)

    (2) part of speech (adjective or participle) and syntax.
    These two distinct categories. And I don’t really understand how these two relate to each other in CGEL.

    One can say that a “participle” is what appears in “verbal” positions/constructions and what appears in verbal positions/constructions must be a participle. This way one can made one category fully subordinate to the other.

    Then if you define what are verbal constructions, you know who are “participles” (but the notion is then fully subordinate to syntax), and if you define what are participles, you know how are “verbal” passives (but the notion is then fully subordinate to lexicon) and if you define both… Well, you either propose a law (that a lexical and syntaxtical category perfectly correspond) that can simply be wrong, or you note that this correspondence mostly works well, but there are some words for which the law does not work well.

    But both adjectives and participles can be used attributively or predicatively!

  192. In other words, I would love to see (in the CGEL) a bit more explicit discussion of these two things:
    (1) the nature of the distinction between the adjectival and participal categories (lexical and syntactical) (2) how syntax relates to lexicon (again in the context of our adjectives).

  193. Sigh. I was seeing a space opera dream (which does not happen often to me) set in a space like in the Little Prince (where “planets” were mere tens meters from each other and I could jump from a planet to planet) and our (I was with some lady) adversaries just blew up a base of some mighty animal-like (bird or frog or both -like) species – and right next to the home of that species! – which clearly meant a war and someone described as “the legendary killer from the [TV] series ‘Avesta’ ” was about to arrive when something woke me up.

    On the other hand, if I did not wake up so abrubtly I would have forgotten everything.

  194. Lars Mathiesen (he/him/his) says

    FWIW, Standard Danish is the proud possessor of a robust neuter noun gender (opposed to the common) that forces agreement in attributive and predicate adjectives. And demonstrative pronouns that have to agree with the referenced noun and forces predicate agreement.

    But only in the singular. There is no fixed relation between noun gender and the form of the noun plural, unlike what seems to be the case in Nynorsk; adjectives and pronouns also have just a single plural form. Maybe the three-gendered dialects do a better job, but I don’t think so.

    (There is a noun class with zero plurals that only contains neuters but the converse does not hold).

  195. David Eddyshaw says

    @drasvi:

    In a way, the distinction that CGEL is drawing is not so much between participles and (other) adjectives but between two different uses of the verb “be”: in the first case, it’s an “auxiliary” forming the passive (though CGEL doesn’t use this terminology) and in the second it’s a copula. Among other differences, in the first use it can form a progressive, and in the second it normally doesn’t, so e.g. “I’m being married” can only be taken as passive, because it’s dynamic (as they mention, this by itself is not an adequate criterion for distinguishing the two constructions, because the passive of stative verbs is also stative anyway, and thus usually doesn’t form a progressive.)

    The dynamic/stative thing pervades the English verbal system; I don’t think that there is anything really similar in Russian (though I know much too little about Russian to say if does something analogous by a quite different grammatical route.)

    you either propose a law (that a lexical and syntactical category perfectly correspond) that can simply be wrong, or you note that this correspondence mostly works well, but there are some words for which the law does not work well

    Part of the excellence of CGEL is that it specifically doesn’t claim that everything is neat and tidy in English grammar, but happily discusses edge cases and peculiarities that don’t fit well into the common patterns. Their discussion on “adjectival passives” is a good example. They don’t claim that any of their tests can actually resolve all ambiguities. However it’s not necessary to do this in order to say that there actually are two different constructions lurking under one surface form: just because there are ambiguous cases, it doesn’t mean that the distinction itself is a mirage. (English gender is still sensibly called “natural” and not “grammatical”, despite the fact that traditionally, ships are “she.”)

  196. David Eddyshaw says

    I think I’ve confused the issue by attributing the dynamic versus stative distinction to the participle/adjective itself rather than to the construction as a whole, which is not what CGEL is doing. That was me, not CGEL.

    I think that this is legitimate, myself, but it’s not a fundamentally different analysis: it’s more of a complementary way of looking at the same phenomena. (I was encouraged to look at it this way by comparison with Kusaal, which has productively formed deverbal adjectives but no participles*, and with classical Greek, in which aspect is clearly a feature not only of participles but even of infinitives.) If you don’t like attributing aspect to English adjectives/participles you’re in very good company: I’m not offended if you atttribute it to the verbs instead …)

    * Like Welsh …

  197. David Eddyshaw says

    Urs Niggli’s dictionary of Burkina Faso Toende Kusaal, which in most respects is pretty good, has an odd frequently-repeated error which evidently arose from automated text substitution* with no subsequent checking of the text. Repeatedly, nouns are tagged as “imperfective aspect.” While I’m sure that this is an error, I did initially wonder whether it was some kind of arcane grammatical point: after all, ordinary non-deverbal nouns are “imperfective aspect”; stative, even …

    * Someone has mechanically replaced something like “inflected forms:” with “imperfective aspect:” throughout, forgetting that it’s not only verbs that inflect.

  198. Is Urs an actual given name or is it short for something?

  199. David Marjanović says

    Urs is a very common first name in German-speaking Switzerland, straight from Latin ursus.

    The rest of the German-speaking area only has Ursula to offer. Compare the Swiss-only Beat & Renat (initial stress of course).

  200. An actual Swiss German given name, etymologically from ursus. It is the masculine version of Ursula. St. Ursus was martyred in Solothurn (of which he is the patron saint) around the year 303.

  201. David Marjanović says

    classical Greek, in which aspect is clearly a feature not only of participles but even of infinitives

    Same in Russian, BTW. Imperatives as well.

  202. David Eddyshaw says

    Same in Russian, BTW

    Kusaal doesn’t have infinitives, but almost all verbs have gerunds, which are quite similar. They are assigned to various noun classes according to fairly reliable rules; interestingly, all gerunds in the -m “liquids, substances, abstractions” class come either from imperfective-only one-aspect verbs or from imperfective forms of two-aspect verbs. There’s an association in there between “imperfective”, a verb aspect, and “mass”, a noun category.

    [It doesn’t carry through completely: two-aspect root-stem verbs usually form their gerunds in -b(ɔ), a class whose only other members are “war”, “porridge” and “soap”, and one or two gerunds in -m have developed concrete senses that can pluralise, like sʋ’ʋlim “possession.”]

    There is also an association between definiteness, a noun-phrase category, and mood, a verb category, with irrealis going with indefinite and indicative going with definite:

    Ba na yɛlif on na niŋ si’em.
    they IRREALIS tell.you she.NOMINALISER IRREALIS do how
    “They’ll tell you what she’ll do.”

    Ba yɛlif on niŋ si’em la.
    they tell.you she.NOMINALISER do how the
    “They’ve told you what she’s done.

  203. David M.: Romance varieties with the neuter typically have both “ambigenericity” (for obvious diachronic reasons: the final /a/ of the neuter plural being identical to feminine singular final /a/, it would be amazing if neuter plurals had not been analyzed as (in some ways) grammatically feminine) and various other markers which are specifically neuter. Standard Romanian has plural -ă for one noun, “ou”, egg (neuter ) (Yes, straight from Latin OVUM/OVA), and in some varieties of non-standard Romanian this ending is found with a great many neuter nouns, to the exclusion of masculine or feminine ones. In Neapolitan, along with ambigenericity-type phenomena with inherited neuters, the definite article and demonstrative, in the singular, trigger gemination of the initial consonant of the following noun if said noun is neuter, and never for masculine and feminine nouns. So: it is a mixed bag.

    On the name “ursula”: its roots go back further than Latin! There is a well-known Late Latin inscription-

    https://edh.ub.uni-heidelberg.de/edh/inschrift/HD077356

    -which refers to a mother named “Artula” and a daughter named “Ursula”: “Artula” being the Gaulish word for “she-bear”, it has been argued that, in the context of language shift from Gaulish to Latin, the daughter was named after the mother.

  204. David Eddyshaw says

    Artie!

  205. Good 2020 discussion of PIE *rkto- at LLog, reminding us that ursus, artula, and arctic are all one. (See also the earliest Medvedic sutras.)

  206. David Marjanović says

    The last comment there links to this LH thread which discusses ursine etymology in great detail (and links back to the LLog thread, not to mention to itself extremely often). Complete with a Rakṣasa-sūtra.

  207. Same in Russian, BTW. Imperatives as well.

    Use of aspectual forms in Russian imperatives is material for more than one thesis.

    Sometimes they work just like any other aspects, namely when you ask to do something habitually (“[always] start your letters with….” has ɪᴘꜰ because habitual) or to keep doing something (“speak slower” is ɪᴘғ * because ᴘʀᴏɢ).

    But when you ask to initiate an action (“jump”! prototypically perfective)….


    * toying with a small cap generator…

  208. not sure if it’s a related phenomenon, but i just saw this sentence here:

    “If the musician is bored, so too will the audience.”

    it rings as wrong to my ear in a similar way to “we could read but aren’t”.

  209. Same here. Great find!

  210. “If the musician is bored, so too will the audience.”

    Interesting. Consider various substitutions that make an acceptable sentence:

    “If the musician gets/becomes/grows bored, so too will the audience.”

    Or even:

    “If the musician waxes/proves/appears/seems/looks/feels/acts/leaves/stays/retires/behaves bored, so too will the audience.”

    it rings as wrong to my ear in a similar way to “we could read but aren’t”.

    Now that’s interesting too!

  211. Not wishing to flog a dead horse, but wouldn’t it be simpler to change it to “we could, but aren’t, reading” ?

  212. That’s just as bad.

  213. January First-of-May says

    That’s just as bad.

    If anything, it’s significantly worse (to me, anyway).

  214. David Eddyshaw says

    I notice a whole paper on this subject in a fairly recent Language:

    https://www.linguisticsociety.org/sites/default/files/e05_97.1Geiger.pdf

    (I’ve kinda given up on Language as a regular thing, on the grounds that too many of the articles seems to be of interest only to Chomskyite initiates, but this one doesn’t seem to fall into that category, anyhow.)

  215. I assemble these three texts for comparison:

    1. But the project was abandoned due to technological challenges and a shortage of funding. / What this means, then, is that there are codes inscribed in ancient manuscripts that we could read but aren’t.
    [From the original post, with the wording of most interest in italics.]

    2. If the musician is bored, so too will the audience.
    [Adduced later by Rozele, and I commented that various substitutions for is make a more acceptable sentence.]

    3. [I]t rings as wrong to my ear in a similar way to “we could read but aren’t”.
    [Rozele’s comment on 2.]

    What interests me now is the matter of saving context. The italicised core of 1, in a different setting:

    1a. While more recent and accessible codes are fascinating reading, there are codes inscribed in ancient manuscripts that we could read but aren’t. Most record ordinary commercial transactions.

    A little clumsy, but far more acceptable now. We can give 2 a saving context also:

    2a. The critics will leave if the performance is substandard. If the musician is bored, so too will the audience.

    A meaning different from the one intended, but an acceptable sentence at least.

    Now, I remarked that 3 was interesting. Why? Because like 1 and 2 it strikes me as borderline at best, combining two constructions that appear incompatible:

    3a. It rings as wrong to my ear as “we could read but aren’t”.
    3b. It rings wrong to my ear in a similar way to “we could read but aren’t”.

    And I cannot immediately find a new context for 3 that will save it. Just saying, is all. But now I see that I was hasty with a certain assumption. In 3 and 3a the as may have a role that I failed to notice, like this:

    3c. It rings as wrong to my ear.

    I had assumed that this was the only normal usage:

    3d. It rings wrong to my ear.

    I prepared this comment before looking at the paper linked by David E, which as I now see deals explicitly with context – in a different way? I’ll read it soon.

  216. In support of my somewhat hasty assumption, ngrams for “it rings true” (common) and “it rings as true” (quite rare; nothing like it found in the full text of OED except something that confirms my take: “1872 … I…heard it ring as true as tested gold.”). And ngrams for “rings wrong” (found) and “rings as wrong” (not found). Indeed, “open-ended” ngrams for “rings as true *” give further support for my assumption about the role of as.

  217. my “rings as wrong” was definitely not the comparative “as [wrong as]”, and may or may not be pure idiosyncrasy. i wouldn’t use “rings as true”, except in the comparative construction, but “rings as” is definitely a cromulent form for me with many other adjectives (though not, i think, with ones that are actually about sound).

    i wonder how other northeastern-u.s. anglophone folks feel about it? it’s quite possible that it works for me because of yiddish “klingt vi…”.

  218. “If the musician is bored, so too will the audience,” just looks like an editing error to me, where “gets” (or equivalent) was changed to “is.”

    However, my reason for commenting on that sentence is actually not linguistic. I just wanted to push back against the idea the sentence expresses—that performers and audiences frequently share an implicit emotional bond during a live performance. I know from experience, that the audience’s perceptions of how a musical performance goes are often starkly at variance with the feelings of the performers. Objectively terrible (and unpleasant for the musicians involved) performances can nonetheless go over really well with audiences, and the reverse happens as well, with the performers feeling that they were really into it but the evoking a lackluster response from the audience.

    There certainly can be situations in which that kind of communion between musicians and audiences can happen, and it is probably easier with smaller groups of musicians—such as solo acts—who can more clearly display their emotional investment. It is part of being human that people very naturally want to latch on to superficial indicators of shared emotional response. However, the outward signs of that investment are also things that can oftenly be relatively easily be faked. Consider Leonard Bernstein’s absurdly over-the-top conducting style, particularly later in his career. Or note that people find sitcom jokes funnier when they are accompanied by a chorus of canned laughter.

  219. David Marjanović says

    i wonder how other northeastern-u.s. anglophone folks feel about it? it’s quite possible that it works for me because of yiddish “klingt vi…”.

    Interesting – I (would) render that as sounds like, not as rings as (which I’ve never encountered either).

    But then, what the telephone does is läuten over here.

  220. David Marjanović says

    I’m in command! I could order this. But I’m not!

    …ordering it, that is. But that’s never said; after a long dramatic pause, the speech continues with “because”.

    (…Rhyme not intended.)

  221. Brett:

    “If the musician is bored, so too will the audience,” just looks like an editing error to me, where “gets” (or equivalent) was changed to “is.”

    Plausibly, in which case all bets are off. But we can’t be sure. Cases like this are common in conversation:

    “I hope you’ve got the tickets, because I don’t.”

    Or more frequently with two speakers:

    “Got time for a cuppa?” “No I don’t. Emails to finish.”
    “Have you got the tickets?” “No, I thought you did.”
    “Audrey Tautou was adorable in Amelie.” “I do! I loved her in Priceless too.”

    I notice this especially across the US–UK+ boundary. But examples like the last were a feature of conversations I had with a highly articulate native speaker (non-US) graduate with a major in English literature.

  222. i think this is a bit different, maybe in a way that’s specific to the verb/situation.

    in the first two exchanges, “[have you] got” is being parsed as “do you have”, which are in fairly free alternation, which allows the object to be dropped. a response with “got”, especially a negative one, would have to be longer and would feel a bit repetitious: “i’ve got time”/”i haven’t got time”.

    the Amelie example feels different, and sounds a bit off to me (though it would fly right by in conversation), in a ‘replying directly to an implicit question’ way (“don’t you think?” or “do you like her?”).

  223. Rozele:

    A bit different but not so easily dismissed – and equally interesting at the contested boundaries of “grammaticality”.

    in the first two exchanges, “[have you] got” is being parsed as “do you have”, which are in fairly free alternation, which allows the object to be dropped.

    If parsing is a matter of finding the grammatical roles of a word or words in a string, then what you give is not an example of parsing. The semantic content is the same in “[have you] got” and “do you have”, though their syntactic structures (and parsings) are at odds. Isn’t that the point with the earlier examples too?

    the Amelie example feels different, and sounds a bit off to me (though it would fly right by in conversation), in a ‘replying directly to an implicit question’ way (“don’t you think?” or “do you like her?”).

    Or the verb root of adorable is taken as presented, and assumed in the reply “I do [adore her]!” Again, I don’t see these as essentially different from the earlier “off” examples. Maybe just a further (or differently positioned) remove from canonically structured discourse, and amply attested.

  224. i meant “parsed” a bit more broadly: ‘structurally understood’ (but not in terms of the specific words’ roles, which i don’t think non-linguist speakers think about unprompted), in this case not as phrasings whose only relationship is being semantically substitutable* but as more and less formal syntactic alternatives for handling a question & answer using “have”.

    and the difference, to me, in the Amelie example is that i don’t see “i do” as having any connection to the preceding statement except a semantic one (created entirely by juxtaposition), while the others here (and in the earlier examples) are all parallel constructions with a shared verb**, whether the parallelism is between clauses or speakers.

    .
    * as “got time [for a cuppa]? / i can’t stop [for refreshment] right now.” would be.
    ** taking the “got” and “do” constructions as interchangeable alternatives using “have”, that require somewhat different sentence structures.

  225. I personally think sentences like “I have already eaten, but he still is” or the Amelie sentence are in this intermediate zone, where the supposed logic of language is unquestionably violated, but where the meaning is clear. To me this logical laxity allows for more sstraightforward and clear language, and is admirable and worth imitating, in speech at least.

  226. @DM (…Rhyme not intended.)

    But those don’t rhyme – at least not in my English.

    ‘Pause’ is a homophone with ‘paws’.

    Vs ‘becos’. Only if I was making some extraordinarily emphatic connection would I drawl ‘becaws’ to that extent.

    I’m afraid the claims here about what is and is not interpretable/parseable have gotten beyond the point of my having reliable intuitions.

    The “…, but he still is.” example I had to do a double-take. Oh: still is eating. Of course I’d get it in context, but it still doesn’t ring right.

  227. Of course I’d get it in context, but it still doesn’t ring right.

    Same here. I’m fascinated by the widely varying acceptability of these sentences among native speakers, and wonder what it means for the confident placement of asterisks by linguists.

  228. The Amelie exchange just sounds like the second person didn’t hear the first properly but got the general meaning. It’s hard for me to imagine they’re actually treating “adorable” as a verb form.

  229. David Marjanović says

    Only if I was making some extraordinarily emphatic connection would I drawl ‘becaws’ to that extent.

    Ah, I figured that would happen here because it’s stressed and prepausal.

  230. I’m fascinated by the widely varying acceptability of these sentences among native speakers, and wonder what it means for the confident placement of asterisks by linguists.

    I’m not native, but still…
    One big problem with the asterisk notation is that it’s too laconic. It encompasses too wide of a range of reactions, from “huh?” to “illogical but kinda neat” (mine) to “of course I’d get it in context but it still doesn’t feel right” (AntC’s).

  231. Exactly, but linguists (at least when I was in grad school, back in the Cretaceous) treated them as a yes-or-no indicator. (I used to torment the professor in the Chomskyite course I was forced to take by saying sentences he was marking unacceptable were fine in my dialect.)

  232. Ah, I figured that would happen here because it’s stressed and prepausal.

    The rhyme works fine for me (American).

  233. David Eddyshaw says

    Exactly, but linguists (at least when I was in grad school, back in the Cretaceous) treated them as a yes-or-no indicator

    This article of Geoffrey Pullum’s (which I strongly suspect that I must have linked to before) is very much concerned with this very same major defect in the Chomskyite/generative systems:

    http://www.lel.ed.ac.uk/~gpullum/Consigning.pdf

    Take the section of my article in which I point out that GES grammars define only a binary grammaticality distinction. They classify everything as either perfect or nonexistent—the only alternative to being grammatically impeccable is being nonlinguistic garbage and not having grammatical properties at all. But in reality, I point out, ungrammaticality seems intuitively to be a matter of degree: some utterances are much more ungrammatical than others. So there is a gulf between the phenomena and the theoretical account.

    All Neeleman says to this, really, is that he doesn’t intend to worry about that. The matter can be treated as relating to performance rather than competence. The judgments of grammaticality people make are a matter of performance, and ‘we expect, irrespective of the nature of the competence grammar, to find variation in grammaticality judgments.’

    (This reply to Neeleman is a good bit more digestible to groundlings like me than the original paper that N was – according to P – inadequately responding to, and conveys the point well enough.)

  234. Still. A bunch of people figured on solving this problem by using a range of numbers between 0 and 1 to assign to acceptability, instead of just 0 and 1. That is still an engineer’s solution. “Acceptability” covers a range of qualitatively different psychological reactions to an utterance, and scalar acceptability or whatever they call it does not address that.

  235. Keith:

    It’s hard for me to imagine they’re actually treating “adorable” as a verb form.

    No one has suggested that.

    David E:

    Pullum is just as militantly binary-thinking, in the domain of prescriptivism versus descriptivism. Any attempt to improve text that goes beyond appeals to grammaticality is to be suppressed with extreme prejudice, it often seems.

  236. David Marjanović says

    “Acceptability” covers a range of qualitatively different psychological reactions to an utterance, and scalar acceptability or whatever they call it does not address that.

    seconded

  237. seconded

    Thirded. (See my preference for scare quotes when writing of the “grammatical”, and my “more acceptable” rather than “more grammatical”.)

  238. David Eddyshaw says

    Pullum is just as militantly binary-thinking, in the domain of prescriptivism versus descriptivism

    Citation needed.

  239. I would cite an exchange at Language Log concerning “that” and “which”, concerning which Pullum has very strong views; but unfortunately the non-Pullum side of the exchange was later discreetly and entirely expunged from the record.

  240. David Eddyshaw says

    My impression of the Mighty P has always been that he’s chill with prescriptivism so long as it knows its place and doesn’t get uppity. Nothing wrong with teaching children how to abide by certain linguistic shibboleths when desirable in order to live long and prosper, so long as you don’t mistake the shibboleths for Rules of Thought or moral imperatives.

  241. My impression of the Mighty P …

    “He’s not the Messiah, he’s a very naughty boy.” But like you I have high respect for CGEL, which has its place on my shelves. I interrogate it often (sometimes critically), in hard copy and in searchable PDF.

  242. David Eddyshaw says

    One of the things I like about CGEL (if not the very thing I like most) is that it is written (mostly) in a way that positively encourages you to read it critically. Extensive discussion of controversial points, frank admission of uncertainty and incompleteness … excellent. No chance at all of mistaking it for an Infallible Guide.

    Incidentally, I was just reading P’s Rarely Pure and Never Simple again

    http://www.lel.ed.ac.uk/~gpullum/LTTCpaper.pdf

    and was struck (again) by

    … as a brief and admittedly very programmatic statement I would say the following. A human language, far from being a tight-knit redundancy-free formal system, is more like a sprawling library of interlocking construction types with a wide range of different productivities, frequencies, and vintages. Like any library, it has contents of very different ages: some constructions are essentially borrowings from centuries ago (as it were, or would to God, or be it ever so humble), while others are brand new (How cool is that! or Oops, my bad!).

    The enormous library in question is capable of being acquired, internalized, and used — to varying degrees for different people — through a process involving massive exposure leading to ingrained familiarity and ultimately a degree of automaticity. The mental storage load is moderated by the sporadic presence of indefinitely large equivalence classes of words or phrases with fully shared syntactic behaviours (this point is made in a little more detail in the final section of Pullum and Scholz 2007).

    This is very broadly the view taken by the proponents of construction grammar: Charles Fillmore, Paul Kay, Arnold Zwicky, Adele Goldberg, and also Ivan Sag and others in the HPSG tradition. There are many open problems about how construction grammar might be made fully explicit (Paul 2002 makes a start on the problem of devising a formalization), but the top-level insight seems to me broadly the right one.

    Developing the construction-grammar view fully is of course a research programme to be worked out over decades. CGEL can be seen as a step in the direction of providing the descriptive basis for such a research programme, and a move away from the kind of theoretical linguistics that leans toward radically selective attention and purely fictive regularity …

    I’d forgotten that he was so positive about Construction Grammar. I agree with him that its approach seems basically right, but that there is an awful lot of work waiting to be done on the detail …

  243. more like a sprawling library of interlocking construction types with a wide range of different productivities, frequencies, and vintages.

    For example: slap in the middle of the Great Vowel Shift, would Generativists be forced to describe English as two languages?

    Presumably speakers from opposite ends of the country could understand each other, though each would be producing a different set of phonemes. Putting all that down to ‘performance factors’ seems preposterous.

  244. “My impression of the Mighty P has always been that he’s chill with prescriptivism so long as it knows its place and doesn’t get uppity.”

    Hm.
    I do think that schooling is not innocent. I think (1) a hour of a child’s time is no less valuable than a hour of an adult’s time (2) school and its endless evaluation of language form the position “correct or not?”, and corrections and grades do affect (or shape) us.

  245. It is not just Pullum of course.
    But “is not ready to comment on school education” is hardly a compliment (from my point of view).

  246. I think (1) a hour of a child’s time is no less valuable than a hour of an adult’s time

    This is one of the main points of Helen DeWitt’s The Last Samurai.

  247. One of the things I like about CGEL (if not the very thing I like most) is that it is written (mostly) in a way that positively encourages you to read it critically.

    Before we get too carried away praising Pullum, remember Huddleston the (equal if not greater) co-editor. Perhaps the dialectic just reflects those editors couldn’t agree — which absolutely deserves admiration.

  248. David Eddyshaw says

    remember Huddleston

    True that. (But I hadn’t forgotten …) CGEL is a work of many hands, in fact.

  249. This top-level insight sounds a bit like an attempt to explain language (rather than a humble model that helps to work with some aspect of it).

  250. David Eddyshaw says

    I don’t think the two are disjoint: pure evidence-collection without any kind of theory about what you’re investigating is impossible. (If you think that that is what you are actually doing, it’s because you are blind to your own unexamined assumptions.) Problems come when the theory starts leading you to spend most of your time in picking your data to support it, and when you refuse to accept that your theory needs modification – or even complete abandonment. But you’ve got to start somewhere. There’s no tabula rasa.

    Theories (however) can be so ill-judged from the outset that they fatally blind you to what’s in front of your very eyes. Chomsky’s various protean theories (before he helpfully divested his programme of all actual potentially refutable content) do this – and this is a good instance. The theory itself makes it impossible to cope in any principled way with the fact that grammaticality is not all-or-nothing.

    Even on the level of simply trying to write a useful descriptive grammar of a single language, it’s not possible to be completely atheoretical. The idea reminds me of the quip that “common sense” is, in point of fact, usually based on quite complex premodern scientific theories, which have been discarded in actual science because they don’t actually work.

    Decent modern descriptive grammars* often pray in aid R M W Dixon’s “Basic Linguistic Theory” for reasons like this.

    https://en.wikipedia.org/wiki/Basic_Linguistic_Theory

    I have some doubts, myself, about the actual existence of this particular unicorn (languages vary too much among themselves) but I understand the point of the idea well enough.

    * i.e. those that you can actually get useful information out of regarding the language in question, as opposed to those written primarily to display the great power and wonderfulness of the One True Programme, and the author’s corresponding suitability for publication, employment and honour.

  251. How does Quirk et al.’s grammar hold up next to CGEL? Is there a significant element of equal but competing approaches, rather than just of one being newer and so more complete?

  252. R M W Dixon’s “Basic Linguistic Theory”

    For all the good that Dixon has done, he is not famous for his self-effacing shyness.
    BLT (the book, not the sandwich) is great for what it is. However, the sad truth is that if you want to delve into the syntax of complex sentences, there is still nothing to guide you usefully that doesn’t derive from the generative tradition. IIRC Haspelmath mentioned years ago that he was going to do one of his treatments of the concept of “consituent”, but that hasn’t happened yet. That, unfortunately, is the state of the art on our side.

  253. David Eddyshaw says

    How does Quirk et al.’s grammar hold up next to CGEL?

    I’m not nearly as familiar with the “other” CGEL (which I also have a copy of, though by no means as well-thumbed.)
    [One of the authors was a friend of a relation of mine …]

    It’s certainly very good indeed, but my (fairly uninformed) opinion is that it’s a good bit less bleeding-edge, as it were, in the sense of being more narrowly descriptive, and less concerned with trying to understand how things work. H & P is more into theoretical issues as such.

    That’s why I tend to look at H & P much more often, really; I’m often looking at it, not so much to find out about a fine point of English grammar, but to see how some real experts have dealt, in a familiar language, with a cross-linguistically common problem in grammatical analysis.

    I reference H & P in my Kusaal grammar several times, not because Kusaal is much like English, but in order to make sure that my own analyses are not getting too far out of step with the sort of thing that the experts have thought possible (and, I hope, to give readers a relatively familiar point of comparison.) As a particular instance, I found H & P’s notion of “catenative” constructions (which does a lot of work in their analyses) very fruitful in getting a handle on the complicated issue of clause subordination in Kusaal. Quirk et al doesn’t really lend itself to that sort of use (and why should it?)

  254. I wonder whether (per contra) the spelling might actually represent dental rather than alveolar stops?

    I wonder if the appearance of childher shows that this verse is not quite traditional Lancashire dialect, but proto-Scouse, which is a mixed Hiberno-Lancashire variety.

    [W]ait, is being an abomination a necessary condition of being eldritch?

    Definitely not. The OED defines eldritch as ‘weird, ghostly, unnatural’ and only then ‘frightful, hideous’. An alternative form is elphrich, and so the word is probably connected with elf, as in Stewart’s 1535 translation of Boethius: “Thair cleithing quhilk [which] wes of elritche hew.” The un-updated OED’s latest quotations are “Truth is appalling and eltrich, as seen By this world’s artificial lamplights.” and “Joy that had something eldritch and unearthly in it.” But in the DSL, we find it as recently as 1992: “A corncrake — that’s what it is. God, that’s rare nowadays. Very few left, I was reading. What a lonely, eldritch kind o a sound it is. Krekk!”

    Of course whether you feel the unearthly, or at any rate unhuman, is abominable depends on your general view of the Other: whether its power to fascinate is science or horror. Elves are about as Other as there can be.

  255. David Eddyshaw says

    The fairy king in the excellent Jonathan Strange and Mr Norrell is pretty damn eldritch, but not at all Lovecraft-abominable … (still a good idea to run away fast, though. The worst thing is if he actually likes you …)

  256. @DE, well, for example: if “classes of words with similar syntactic behaviour” are due to what he said, so is syntax as a whole (and grammar as we describe it as a whole). Which implies two very bold statements:

    – it is possible to communicate without what we know as grammar, the only problem with such communication is what he said (maybe, but could he, please, demostrate what such communication could look like with a conlang?)
    – he knows why we use grammar.

  257. I left out the OED final dates for eldritch: they are both in the 1860s.

  258. David Eddyshaw says

    it is possible to communicate without what we know as grammar

    Sure we could, if only our poor monkey brains had the capacity.

    The idea (I think) is that “syntax”, far from being some genetically determined thing mysteriously lodged in our brains, is actually a bunch of heuristics we’ve come up with so we don’t have to memorise every single utterance separately, which we can then leverage to make even more utterances (should we be so disposed.) Different people may well use different heuristics to achieve the same ends, and even individuals may have a range of heuristics that they use at different times or for different exact purposes.

    If so, you’d expect these heurisitcs, for the most part, to be instances of the more general capacities that we have for pattern recognition and so forth. To me, at any rate, that seems very plausible.

    You’d probably still be left with some language-specific stuff; personally, I’d expect that to be more at the phonology level. The human ability to segment speech into phonemes, for example, is positively uncanny, given the fact that in terms of actual sound, they don’t exist …

  259. David Eddyshaw says

    Conlang to demonstrate, as requested:

    A: “I’m going to Wigan on Thursday.”
    Ba: “Do you want any ice cream, or would you rather have an elephant?”
    Ca: “I’ve never really felt that Chomsky’s work was the breakthrough that his acolytes claim.”
    Da “Bang!”
    ….

    and so forth. (The general principles are simple, of course, and clear, I hope.)
    The only real problem with this conlang is learnability*, but I think that is true of Lojban, too.

    * Well, OK, the dictionary is very expensive – and even the pocket edition is quite heavy.
    But it’s worth it For Science!

  260. Problem is, a child would never get past the impossible-to-answer Ba.

  261. Ca

    @Yyy (but I am afraid, @ too must be incorporated. And maybe even “drasvi says
    November 24, 2022 at 8:11 pm” etc.).

    And I don’t know what to do about click to edit 13 minutes and 58 seconds…

  262. David Eddyshaw says

    You ChomskYites and your recursion combinators!

  263. DE, I’m not sure how coining/aquisition etc. are going to work. Also numbers (and also quotemarks etc.) illustrate the problem: we actually can name indifinitely many numbers.
    Yes, what Ch. calls the Galilean challenge. You devised a conlang that does not have this property.

    It seems Pullum means that in natural languages/idiolects (with a given size of the lexicon) “classes” arise for only one reason. This “only” is going to bother me irrespectively of what he means, but if* without classes we actually have no grammar, a valid example would be a conlang whose lexicon has the same size as yours, and whose expressive power is also the same. Just without grammar.

    And I just don’t understand how such a thing could work….


    * we have words #1, #2, #3 and constructions xxxxx ____ xxx, yyyy _____, zzz _____ zz z.
    And we observe two properties

    – we can fill in the blank in xxxxx ____ xxx with any of words 1, 2, 3. There is more than one word that can be used in this construction, in other words.
    – if we can fill in the blank xxxxx ____ xxx with some word, we can also fill in blanks in yyyy _____, zzz _____ zz z with the same word.

    I think P has in mind the second property or both of them. When both are true, we have a table 3×3 where all possible combination work. When none is true I just don’t see what is left of our grammar.
    But does the former property alone makes words 1, 2, 3 a “class” in his understanding?

  264. Sorry Stu:(

  265. In other words, I understand Pullum as saying that when a language instead of tables m×n has tables 1×n it is inconvenient. I’m saying that it is inconvenient for many reasons (why call one of them the main reason and dismiss others?) and that 1×n seems to create his classes as well (and I can’t imagine a 1×1 language).

  266. David Eddyshaw says

    There’s nothing in (what I take to be, possibly wrongly) to be the Pullum Programme that precludes really any kind of mental construct getting incorporated into language as a subsystem.

    In fact, with numerals, it seems impossible to me on first principles that their manifestations in language are primary at all: they arise from other parts of human cognition, and get language labels slapped on them secondarily. (Bear in mind that speakers of Australian languages with “no words” for numbers after “three” or so may still be able to count perfectly well using gestures.)

    More broadly, all sorts of traditional syntactic apparatus can find a place in the New Order: it’s not that it’s forbidden from existing at all, more that it’s demoted from being some sort of built-in “optimal”-because-ANC-says-so design feature of the human mind to a set of mental constructs on a level with how we recognise and manipulate patterns of all kinds, not just of words.

    The Construction Grammar people call pretty much anything a linguistic “construction” if it links form and meaning*; although the simplest case is a morpheme, much more complicated things, even word order, can qualify. The tricky bit is saying how it all fits together. Very tricky. The task is much harder than Chomskyite projects of describing Platonic-ideal pretend “language(s)”, not least because the whole point is describing all of a real language As She is Spoke (including things like “As She is Spoke.”) Hence Pullum’s ETA of “decades.”

    You devised a conlang that does not have this property

    Not a bit of it. It has a unique word for every number, and indeed for anything you might say about numbers. It’s just that these words don’t actually have any recognisable relationship to one another: there’s no syntax, so they can’t have any.

    I’m not sure how coining/acquisition etc. are going to work.

    It can’t work. The language is not learnable by human beings. That doesn’t mean it’s not a language at all: it’s just that the constructions of which it is made, linking form and meaning, are all completely unanalysable into smaller constructions.

    Obviously no child could learn a language like that, although in the very early “holophrase” stage, that is exactly what a child takes language to be. But once it progresses beyond that, it begins to start spotting patterns and regularities in the speech it hears and analysing them into more than one construction apiece, and eventually using this to produce new utterances never heard before by creatively recombining constructions.

    It can do that exactly because no human language can ever consist just of strings of unanalysable monolithic constructions: the parents themselves could never have learnt such a language. But there is no point that this analytical process must get to: some constructions actually never get analysed down to verb/subject/whatever level, and others are only partly amenable to analysis: that’s just the way they are. (The Construction Grammar people have lots of pet exhibits of such things in English.)

    * I think this is actually a weak spot in their system(s). As a good (later-period) Wittgensteinian, I am very wary of ideas that seem to imagine that the link between form and meaning is at all simple, even in principle, let alone in messy real-language/real-world cases.

  267. DE, are you sure you want to get the definition of language that expensive? Usually, the way to understand something is not to take the broadest possible view, but to limit oneself to something manageable. Like free fall motion without air friction. But there is air friction! Well, too bad, let’s first figure out what happens when we ignore this pestilence. But the real world! We’ll get to it in due time. First, let’s understand at least something.

    By the way, would you consider traffic signs a language? I bet most linguists wouldn’t and for a good reason.

  268. By the way, would you consider traffic signs a language? I bet most linguists wouldn’t and for a good reason.
    Why not? They form a system for communication, they have categories / classes (e.g. red circles mean prohibition) and syntax (placing something inside the red circle indicates what is prohibited, rectangular tables below limit or specify the validity of a sign), etc. The system is limited in what it can (and is meant to) express, but I don’t see why it shouldn’t be called a language.

  269. David Eddyshaw says

    are you sure you want to get the definition of language that expensive?

    Sure, why not?

    You can indeed then go on to take a particular interest in actually-existing human languages, with interesting contingent properties like being extremely analysable into smaller chunks, and having a physiologically fascinating basis in sound perception and production (usually.)

    But the advantage of having a very broad a priori conception of “language” is that when you do study such languages, you haven’t already smuggled in limitations on what you think that they can possibly be like; limitations which in fact flow from your definition of “language” rather than observing how people actually communicate. And your discoveries about human language will be real discoveries about real-word contingent facts, rather than mere reflections of your own hidden assumptions.

  270. Why not?

    In the first instance, people are usually not allowed to put up they own traffic signs. I also have a bit of a problem with imagining traffic signs’ dialog.

    DE, the problem is not that people cannot limit themselves to study a paricular language or a group of languages. In these cases the general definition what is language is basically irrelevant. If someone doesn’t want to call whatever someone else is studying “language” there’s not much of a problem. I mean, grants, academic positions etc. will be a problem, but not a scientific one. The real question is what can be lerned cross-linguistically and there I think some limitations and simplifications are inevitable. All or almost all languages have something recognizable as syntax. Fine, let’s study syntax and worry about counterexamples and edge cases later.

  271. @DE, I just think it is not what Pullum means…. I think he means there could be a language with vocabulary as large as ours with the same expressive power as our.

  272. David Eddyshaw says

    The real question is what can be learned cross-linguistically and there I think some limitations and simplifications are inevitable

    Exactly so: that is the veritable origin story of syntax. We need (and create) syntax so that we can learn a language at all, and then go on to use it creatively; the stark limitations on our memory capacity alone, let alone anything else, make this inevitable.

    Syntax is the limitations and simplifications we need to be able to cope with the task. (If Martians were cleverer than us, maybe their languages would need less syntax …)

    Fine, let’s study syntax and worry about counterexamples and edge cases later

    The problem with that is that it can readily lead to thinking of regular syntax as primary and the edge cases as deviant. The basic idea of Construction Grammar (I think) is that it’s fundamentally deviance all the way down, and the regularities are created by induction over the deviants and subsequent analogical levelling. That doesn’t mean that the regularities are imaginary or unimportant (grammar-writing would be kinda pointless if it were so) but it does mean that Chomsky-like speculations about the actual nature of human language capacity are chasing the wind: it doesn’t arise from some genetically inbuilt perfectly-optimal facility corrupted by mere “performance” at all, and any attempts to force it into that mould will prove sterile (as indeed, they clearly have done.)

    The idea that the regularity really is the “real” grammar is also hard to square with how children really learn language, and probably also with neurolinguistic findings (not that I know much about that, and I suspect that in any case a lot of the data get overinterpreted to fit whatever mental model of “language” the researchers began with.)

    In practical terms, it means that in describing an individual language it may not be a good idea to postpone looking in depth at edge cases and exceptions, or confine them safely to an appendix. The edge cases may well have something important to teach you about the underlying way the language works. This would be the more important the more “exotic” the language appears to the researcher, because the researcher, try as they might, is going to be preferentially looking for regularities of a kind that appear in languages they are already familiar with. But the language being looked at may keep its deviancies in a quite different part of the grammar.

  273. David Eddyshaw says

    Ideophones are a case in point; they are usually treated very cursorily and inadequately in grammars of African languages (including mine, alas) because they don’t correspond nicely to familiar SAE categories:

    Li anɛ pielig fass fass.
    “It’s very white.”

    Li anɛ sabilɩg zim zim.
    “It’s very black.”

    Li anɛ zin’a wim wim.
    “It’s very red.”

    So what’s the Kusaal word for “very”?
    Ideophones are far from marginal in many African languages (Kusaal is actually fairly stingy with them, comparatively.)

  274. As for your conlang, I don’t see how a child can understand that Ba is “Do you want any ice cream, or would you rather have an elephant?”, or how you can inform someone (who does not know that already!) that “drasvi is a natural phenomenon observed in Russia”.

    they arise from other parts of human cognition, and get language labels slapped on them secondarily.

    I mean, your conglang is incompatible with mathematical notation or, say quotemarks. When we use quotemarks – or a blockquote like here – we mark a fragment of text by specific means. And can a speaker of such a langauge run a short computer program in her mind, as you do when you write them?

    Your conlang looks like telepathic communication. With a difference. Imagine a person who can just upload her mental state in your mind. But your conlang contains requests and questions – quite specific mental states… Can such a person think “328782”? Or “328782 – 234 = 328548”?

    It is a very productive example leading to many interesting ideas, and it inspires Chomsky too. But I don’t think it can work as a [counter-]example for a specific syntactical point. E.g. if we consider a smaller conlang, designed like yours but with vocabulary similar to that of your normal idiolect, among many questions to such a language there will be expressive power and how it can be acquired.

  275. David Eddyshaw says

    I don’t see how a child can understand that

    As I say, it can’t. The language is not learnable by humans (or possibly by anybody, short of godhood.)

    My remark about Lojban was a joke. Probably …

    What I do think is that my impossible conlang is a limit case of a sequence of actual possible languages where successive languages in the sequence put more and more into the lexical meaning of individual formally unanalysable words and less and less into compositional syntax. Real languages do actually vary along this axis, at least in certain domains; Saussure himself pointed out that German vocabulary is (on average) more “motivated” than French, in the sense that words are more often analysable into their parts: einäugig versus borgne

  276. Stu Clayton says

    Saussure himself pointed out that German vocabulary is (on average) more “motivated” than French, in the sense that words are more often analysable into their parts: einäugig versus borgne

    Seems as if “motivated” is being used here to politely suggest “elaborate”. Where one Lego brick suffices for a French notion, you need a whole set of them to build a German equivalent.

    Finalement, tout le monde peut faire du foin, du moment que le ciel fait pousser l’herbe.

    Perhaps more straightforwardly:

    Trois jours après la mort, les cheveux et les ongles continuent à pousser mais les appels téléphoniques se font plus rares.

  277. @DE, I just think it is not what Pullum meant….

    What bothers me is the tone: it sounds to me not as one more possible model or angle of looking at language but as an explanation what language is and why. This tone makes me jumpy: I want more humble linguistics:)
    His explanation of existence of classes seems to confirm my suspicion. It looks like he wants to discourage people from following the one true way… and offers another true way.

  278. Apart of your conlang we can consider a language that does have grammar, but whose words can appear at ANY position. The Cat loves Masha. The Cat Mashas Love. Masha Thes Cat Love.

  279. David Eddyshaw says

    Sounds like Warlpiri.

  280. Stu Clayton says

    Sez here that Warlpiri has a “perlative”. What the dickens is that ? A special case for talking about oysters ? A timorous comparative ?

  281. ‘Along’. It’s just another thing treated as a case where we would use a preposition.

  282. Stu Clayton says

    Aha. Another case of misleading nomenclature. “Along” as in “along the path” ?

    DeWiki says:

    #
    Perlativ: Bewegung durch etwas
    “durch das Haus hindurch”

    [Tocharisch, Grönländisch, Warlpiri]
    #

  283. Stu Clayton says

    How often does it happen that an expression pattern in a language is given a grammatical name, although no one is sure of what it means ? Is there an ignoramus case ?

  284. I found that I can’t literally translate по диагонали.

    It is not diagonally, по is along/across and against the surface. But “along the diagonal” is вдоль диагонали.

    По appears where English has “down” as in “down the road”.

  285. And of course по-английски, på engelsk:)

  286. Stu Clayton says

    I found that I can’t literally translate по диагонали

    Why do you want to “literally translate” it, whatever that means ? Perhaps it’s not surprising, given that English is not one of your native languages.

    You yourself wrote at that thread: “In Russian it is just an idiomatic way to say ‘inattentively’. It can mean skipping a couple of pages here and there, for example”, and D.O. wrote: “It’s just an expression for scanning or glancing over the text.”

    That should be the end of the story, unless you’re not sure what “glance over” means.

  287. Russian grammarians now borrow English names.
    Thus we have безличный “impersonal”, неопределённо-личный “indefenitely-personal” and impersonal.

  288. David Eddyshaw says

    @Stu:

    I think drasvi is deploying Russian perlatives, as proof-of-concept.

  289. Stu Clayton says

    @David E: proof-of-concept

    What concept is being proved ? The concept of assigning a grammatical category to something not understood ? The concept of “along” ? Humble linguistics is expected, but not proffered.

    At the moment, the use of “perlative case” has been explained in two different ways: as meaning “along”, and as meaning “through”. My question was about that. Until somebody says straight out “nobody knows”, glossy-eyed diversionary tactics will be rejected.

  290. David Eddyshaw says

    “Along” and “through” are merely complementary manifestations on our earthly plane of the One True Platonic Perlative.

  291. Stu Clayton says

    Form is emptiness, emptiness is bad form. I believe I would prefer an echo chamber to a hall of mirrors. A good five-cent cigar would be better than either, humble-wise.

  292. Why not?

    You mustn’t mind Stu; “contrarian” is his middle name. He’s like a Zen master who sneaks up to you with a stick: beware!

  293. Stu Clayton says

    Speak in riddles and carry a big stick. The essence of Teddy Zen ! No sneaking, though.

    The disciple has to want the stick. Many people don’t know what they want, so the master reminds them.

  294. Our Old Irish professor did threaten us with a stick. But never used it.

    (namely she said that before students were seduced and lured into doing something Celtic, but now Celtic studies will work as a Buddhist monastery where they beat everyone with sticks…)

  295. @LH, no problem here. Stu’s contrarianism is partly similar to that of “mathematical” subculture here, and thus familiar (partly, again).

    As for me and contrarianism: when something is both true and absurd, it is funny. Funnier than false and against the common sense or true and known.

  296. In the first instance, people are usually not allowed to put up they own traffic signs. I also have a bit of a problem with imagining traffic signs’ dialog.
    For me, any system that conveys meaning by the combination of arbitrary signs in accordance with a set of rules is a language. Everyone being allowed to use it or bidirectional communication are not necessary parts of the definition for me. Even for spoken human languages, we have examples for languages not everyone is allowed to use (languages that can be used only by those initiated in a cult, languages limited to being spoken by one sex). And I have seen cases where people put up unauthorized traffic signs in neighbourhoods in order to slow down or keep out traffic.

  297. David Marjanović says

    Half a year ago I came across a conference abstract that said the Germanic cognate of по is the b- in bring and maybe a handful of other verbs.

  298. David Eddyshaw says

    Wiktionary, bless its cotton socks, links the English “bring” with the Lithuanian branktas “whiffletree” …

    https://en.wiktionary.org/wiki/bring#Etymology_1

    (Astonishingly, Welsh hebrwng “iead”, really is cognate, per GPC, fishy though the comparison looks: the first bit is *sem- “one/together.”)

  299. David Marjanović says

    …and a whiffletree is a whippletree, which is “a wooden crossbar for a plough or carriage, pivoted in the middle, from which traces are fastened to a draught animal.” OK then.

  300. Ah, “perlative”. I love all of those -lative and -essive case terms. So pellucid in their meanings. I have only a few regrets in this domain, including:

    • Far too little mention of the “circumlative”. It would be best stressed antepenultimate, though OED gives antepenultimate and penultimate for “translative” – also a case in the “lative” series. OED has only an obsolete verb “circumlate”: “transitive. To carry or bring round; spec. to turn or ‘fetch’ round (a limb). Also intransitive”, and derivatives “circumlation” and “circumlatory”. One very rare mention at the Wiktionary entry for Old Armenian “զ-“:

    7. (with instrumental) round, around, about (this usage was called “circumlative case” by older grammarians)
    շուրջ զքաղաքաւն էին ― šurǰ zkʿałakʿawn ēin ― they were around the town
    արկանել զնովաւ ― arkanel znovaw ― to put on him
    զայնու ժամանակաւ ― zaynu žamanakaw ― around that time
    զհասարակ գիշերաւ, զմէջ գիշերաւ ― zhasarak gišeraw, zmēǰ gišeraw ― around midnight
    զինն ժամու ― zinn žamu ― around 9 o’clock
    անկաւ զպարանոցաւ նորա ― ankaw zparanocʿaw nora ― fell around his neck

    • The illative might sometimes be taken in its other meaning “concerned with inference”; nothing much can be done about that.

    • I’d like to see “pellative” as a variant of “perlative”, just as “pellucid” (beloved word) can appear instead as “perlucid”. Similarly, I want to see “perfer” (infer : illative :: perfer : perlative) appearing somewhere – other than in (perhaps someday) Lyre’s Dictionary.

  301. Furthermore, under Prolative case, I read that “In grammar, the prolative case, also called the vialis case, prosecutive case, traversal case, mediative case, or translative case, is a grammatical case of a noun or pronoun that has the basic meaning of ‘by way of’ or ‘via’.” As they explain it, the meanings are actually more varied than in that simple explanation.

  302. Furthermore, under Prolative case, …

    Heh. OED hasn’t caught up there either. For “prolative” in grammar, nothing to do with case:

    2. Grammar. Having the function of extending or completing a predication.

    1965 Harvard Stud. Classical Philol. 69 39 Tetigisse and inseruisse are prolative infinitives after cupiam.
    1997 S. J. Harrison Aeneid 10 82 In poetry this ‘prolative’ infinitive replaces the prosaic gerundive after nouns expressing wish, desire, and other verbal notions.

  303. Can add nothing,.apart of maybe Russian zaplativ “having paid” (and pereplativ “having overpaid”, doplativ “having paid an additional sum” etc….)

  304. Those are not noun cases.

  305. But:
    – their right part is -lativ (Russian ablativ “ablativus” does not have -e)
    – their left part is some prefix….

    Which is how circumlative etc. are formed. The -p- in the middle is a minor detail.

    P.S. just fooling around

  306. Actually I’m imagining some provincial Russian boy, taught by a French teacher some 200 years ago, and looking at all those French -tif terms as something deeply foreign. This joke (reinterpreting Russian деепричастия (adverbial participles?) like поколотив or озолотив as Latin case names) makes me think of that time.

  307. January First-of-May says

    As I say, it can’t. The language is not learnable by humans (or possibly by anybody, short of godhood.)

    I do recall that a certain Ireneo Funes has proposed a very conceptually similar project…

    Regarding numbers, in particular, a quick complexity argument proves that, in the general case, you can’t really do better (beyond a constant factor) than just spelling out the numbers in your preferred base.
    In other words, perhaps your favorite conlang has a very short term for the number 18,446,744,073,709,551,616 [insert your own favorite long number here; I used one that actually does have a very short name in some conlangs], but most other numbers of similar size would by necessity have long terms.

    (In practice most natlangs end up introducing structural bits like “septemoctogintillion” that ultimately result in an extra logarithmic multiplier. But in a conlang it’s perfectly possible to get away with RFC 2550-style shenanigans that can provide structure while staying within the constant factor.)

  308. David Eddyshaw says

    a certain Ireneo Funes has proposed a very conceptually similar project

    Yes; I may call the language Borges (or possibly Borgne.)

  309. Stu Clayton says

    Far too little mention of the “circumlative”.

    And of the circumlocutative, although it is omnipresent.

  310. “Halte! Le mot de passe?”

    “Les borges sont borgnes ce soir.”

  311. Hans, obviously any definition is somewhat arbitrary, but as I tried to say if traffic signs are included in the definition of language it becomes harder to say something useful about languages in general. Where I live people put up “traffic signs” all the time. Most frequent ones are baloons to indicate a party or a child with a flag to slow down. Maybe I should revise my objection 🙂

  312. Everyone who lived in 1990s Russia knows that there is always a special case for handling money.

  313. Ah, yes. Briefcase (nuclear), attaché case…

  314. J1M, I suspect this language has infinitely many phonemes…

  315. David Eddyshaw says

    Not at all. That would be silly.
    However, some of the words are quite long.

    (There is not time to utter the word for “ice cream sundae” before the heat death of the universe.* But we are not concerned with mere performance here.)

    * Unless you gabble.

  316. @DE, I understood Pullum differently, as speaking about a language with vocabulary of the same size as yours and the same expressive power as yours – but without word classes.

    But you tried to construct an extreme example… Should not we, for consistency, also avoid phonology? Else it is vulnerable : phonemes can become morphemes.

  317. David Eddyshaw says

    No, I don’t think that’s what Pullum means at all; he explicitly says that the classes are needed to moderate the demand on memory:

    The mental storage load is moderated by the sporadic presence of indefinitely large equivalence classes of words or phrases with fully shared syntactic behaviours

    My conlang is merely an attempt at sketching the kind of language that might be possible if (per impossibile) this constraint did not exist at all.

    Incidentally, I don’t think that Pullum is by any means a paid-up Construction Grammar Person; but (like me) he feels that their basic approach is probably along the right lines, but needs a huge amount of fleshing out before it can really produce adequate full-dress language descriptions.

    I have the impression that most CG people wouldn’t disagree, at that … it’s very much a work in progress, but, unlike Chomsky’s Program (or whatever it is), it’s actually progressing

    The Chomsky thing is perfect instance of what Imre Lakatos called a “Degenerating Research Programme:

    https://plato.stanford.edu/entries/lakatos/

    [The cheeky characterisation as a “Research Program” is ANC’s own]

  318. @DE I understand Pullum as speaking about a language whose vocabulary has the same size as yours.
    Maybe I am wrong.
    But whether this conlang has a vocabulary “of the same size” or much larger, there are many issues with such a conlang.

  319. David Eddyshaw says

    I understand Pullum as speaking about a language whose vocabulary has the same size as yours

    Sure: but he wasn’t speaking about a language with no syntax. That was me (in response to your request …) Pullum is innocent. Innocent, I tell you!

    Construction Grammar does not picture language as being without syntax. It’s more that it conceives syntax as arising in a sort of bottom-up manner by induction over concrete instances, whereas Chomsky devotees think of it more as top-down, magically imposing structure on all potential “grammatical” utterances.

    As a Nominalist (in the mediaeval sense) I’m philosophically a lot more comfortable with this way of thinking about syntax. “No ideas but in things”, as William Carlos Williams says: in this case, “No syntax but in actual utterances.” As I’ve said before, in my view, syntactic rules are not real: you can happily swap one set for another if it accounts for the data just as well and you find it more aesthetically satisfying. There’s never only one “correct” analysis.

  320. But his words imply that there is only one problem with a language without classes. That’s where I object: I think there are many problems.

    I know two ways to destroy class distinctions:
    N classes of size 1 (every word can appear in exactly one construction) or 1 class of size N (every word can be used in any construction). N is the size of our vocabulary.

  321. David Eddyshaw says

    But his words imply that there is only one problem with a language without classes

    I don’t think that they do: all he’s doing is saying that the existence of classes (which he’s taking for granted) moderates the demand on sheer memory.

    The question of in what way mental classes are needed to comprehend or interact with the world, and what we actually mean by such “classes”, is much more fundamental than anything addressed by linguistics as such: it belongs with the philosophers, not the linguists. When my hero St Ludwig addresses this (which he does, a lot) he’s not really talking about anything linguistic (in the usual technical sense of the term) at all. The characterisation of Anglophone 20th-century philosophy as having taken a “linguistic” turn is misleading, I think: for the most part, it’s only really “linguistic” inasmuch as, as Sweeney says: “I gotta use words when I talk to you.”

  322. @DE, if so, I have no objections (apart of technical maybe). I thought his exact goal is discouraging other approaches.

  323. The characterisation of Anglophone 20th-century philosophy as having taken a “linguistic” turn is misleading, I think: for the most part, it’s only really “linguistic” inasmuch as, as Sweeney says: “I gotta use words when I talk to you.”

    If only, Eddyshaw Agonistes. The linguistic turn – issuing in linguisticism, see below – was all too real (sic) and radical, and it takes its toll even now. I’m with John Heil. Worth quoting at some length (my bold, at the end):

    In an effort to avoid messy ontological considerations, philosophers in the twentieth century have engaged in what, in polite circles, is called “semantic ascent.” Substantive ontological issues are transmogrified into issues concerning language and its application. In this way, talk of states of mind, for instance, is replaced by talk of mental attributions; talk of properties is replaced by talk of predicates; talk of causation is replaced by talk of theories; and talk of objects or states of affairs is replaced by talk of statements, propositions, or sentences.
    Linguisticism – our preferred label for this approach – regards ontology with suspicion. Ontology is unwholesome, dubious, something to be avoided or, if that is inconvenient or impossible, to be minimized. Linguisticism finds expression in the quasi-technical shift from the “material mode” to the “formal mode,” and is linked to the old idea that reduction requires entailment or translation. This has allowed philosophers to argue that, because physical object statements do not entail statements about (actual and possible) sense experiences, physical objects are not reducible to (actual and possible) sense experiences, although the “information content” of physical object statements is exhausted by statements about actual and possible sense experiences (see Martin 1997, 214–15).
    There is a natural bond between linguisticism and species of antirealism, and indeed these can shade off imperceptibly into one another. The flight from ontology, however, is scarcely more than an institutionalized form of repression. If we replace talk of properties with talk of predicates, for instance, we are left with the question of the ontological status of predicates. If properties strike you as odd or ungainly, surely predicates are worse. Are predicates concrete particulars? Classes of particulars? Abstracta? In comparison, being square and having mass 9.11(10)^–28 g (the mass of an electron) seem utterly transparent.
    Linguisticism does not succeed in replacing or eliminating ontology, but only diverting attention and postponing the hard questions. The mistake is to imagine that it is philosophically innocent.
    [CB Martin and John Heil, “The ontological turn” Midwest Studies in Philosophy, 23 (1999), p. 36]

    In philosophy of mind over the last three decades there has been a great deal of analysis founded on logical and other species of possibility. Squarely in the “linguisticist” camp (in the Heil sense, but my own diagnosis here), it is technically brilliant and probably immune against refutation on its own ground. But is that ground the proper arena for establishing facts about mind in the world, or about anything else in the world? I think not.

  324. immune against refutation on its own ground. But is that ground the proper arena for establishing facts about mind in the world,

    You don’t need to loiter round here for long to become deeply suspicious of claims there are ‘facts about mind’ (or even facts in the world) that aren’t so much mediated by the language they’re expressed in (and its worldview) that there’s no independent way to ‘establish’ them in the sense you’re striving for.

    [Speaking as a fully paid-up fan of late Wittgenstein.]

    I get your point about “immune against refutation”; but any other predetermined framework (such as Logical Positivism/sense data) turns out to be equally metaphysically immune.

    The Hattery I find invaluable for substantiating comparativity on a day-to-day human level rather than the Philosophers’ too-abstract Platonic ideal.

  325. The mistake is to imagine that it is philosophically innocent.

    Eddyshaw is clearly guilty, I’ll readily confess to being a conspirator. Pullum and Huddleston also. Are you seriously suggesting anyone round here _doesn’t_ acknowledge a predisposition?

    As somebody said up-thread (or up another current thread), it’s those claiming to have no predisposition to beware of as lacking self-awareness.

  326. (This is Language Hat not Philosophy/Epistemology Hat; otherwise I’d be inclined to reduce Heil et al to mincemeat. Good grief such naivety! – about how science works in general, let alone Philosophy or Linguistics.)

    Oh and before you start telling me about Heil’s academic chops: they’re as nothing to ANC’s – which also impresses no-one round here.

  327. Cute!

  328. David Eddyshaw says

    I had hoped to guard against this misinterpretation with my parenthetical “in the usual technical sense of the term”; perhaps I should have bolded it, or something.

    Wittgenstein (not really an “analytic philosopher” at all, though I presume he falls under Heil’s anathema), as Hat Himself once pointed out, has nothing of any significance to say to a linguist qua linguist: he just isn’t concerned with that at all. Indeed, about the only “analytic” philosopher I can think of who did have anything much to say of interest to a working-stiff linguist was Grice, and even his work is quite independent of any one individual language. (I admit that e.g. Austin tried, but entertaining as his work is, I don’t think it really led anywhere, either in philosophy or linguistics.)

    “The mistake is to imagine that it is philosophically innocent” is an interesting turn of phrase. While I appreciate that “innocent” is here used in a technical sense, even so …. but (like AntC) I actually glory in the implied guilt: Proud to be an antirealist! Proud, I tell you!

    “Immune against refutation on its own ground”: well, yes. I have not, so far, heard of an experiment which could refute Platonism (much less, verify Logical Positivism.) If metaphysical theories were susceptible of refutation they would be physics. That’s not the arena where the game is played.

  329. I love these ontologico-philosophical exchanges, and find that what I used to think were fairly well established conceptions of the world become more and more obviously ignorant misconceptions, and since I no longer know what to think, I wave with pride my banner of Total Ignorance. I, at least, am not a criminal!

  330. David Eddyshaw says

    Among the many frankly ludicrous claims in the almost psychotically hagiographic WP page on Chomsky is that he is “a major figure in analytic philosophy.”

    The kindest explanation is that the author of the shameless puffery has never got round to actually clicking the handy link to “Analytic Philosophy” provided there.

  331. David Marjanović says

    If metaphysical theories were susceptible of refutation they would be physics.

    Even these?

  332. David Eddyshaw says

    Sure. If you could refute Christianity by pointing to contradictions in the Bible, it would be physics. (Also, false.)

    [The particular one flagged up there comes up quite a bit in actual Christian circles (usually as a discussion of the propriety of “putting out a fleece”, if you’re interested in googling not-very-illuminating intra-Christian arguments on the issue: I wouldn’t recommend the exercise.) My own feeling is that the “yes” answer is based on grievous misunderstanding of what the texts in question are actually about. YM, however, MV. However, there are much more clearcut contradictions than that, but I suppose those are among the previous 147.]

  333. David Marjanović says

    I don’t mean the contradictions; I mean the stories in themselves. (I was just too lazy to link to them individually.) They describe tests of metaphysical theories, don’t they?

    there are much more clearcut contradictions than that

    Oh yes, but, again, that’s beside my point.

  334. David Eddyshaw says

    They describe tests of metaphysical theories, don’t they?

    Well, no, I don’t think they do. The existence of God is taken entirely for granted throughout the Bible, and not argued for at all.

    If they are tests of anything about God, it is of the reliability of God; this is why the idea is problematic, in fact. Christians who disagree with me would say that that is not in fact what Gideon (say) was up to: this reflects his lack of faith in himself rather than in God; they would argue that such behaviour is OK if done in that spirit.

    In fact all three stories can be pretty naturally interpreted as pleas to God to help out. They’re not anything like testing a hypothesis.

Speak Your Mind

*