Njideka Agbo and Oreoritse Tariemi write about some unusual writing systems of Nigeria, starting with Nsibidi (which I posted briefly about here):
Conceived by the Ekoi (Ejagham) people of Cross River state in Nigeria, and used by exclusive societies for men of power and authority including the Ekpe Leopard Secret Society, before extending to other regions of Eastern Nigeria, Nsibidi’s ideograms share similarities with the hieroglyphs.
This exclusive form of writing fascinated the colonialists such that Author P.A Talbot described it as “a kind of primitive secret writing” which was communicated on “cut or painted on split palm stems”.
Despite its popularity, Nsibidi began to dwindle from 1900 until the final adoption of the Onwu Alphabet in 1961.
They go on to talk about Ajami (see this LH post) and Ńdébé (see this post), and then a new one:
[Chief Tolúlàṣẹ] Ògúntósìn has designed what he calls the Odùduwà alphabet (the talking alphabet). Speaking to Global Voices, he said that he was inspired to write the 25 alphabet script after he visited his ancestral shrine in Badagry, a historic town in Lagos, Nigeria. Soon after this journey, he began to have constant dreams where he visited the sun and was “shown the alphabet in the form of lightning.”
They end with this utopian suggestion:
[Kọ́lá] Túbọ̀sún is of the opinion that only a pan-african script that can transcend all Nigerian languages will effectively work.
“The biggest problem I see with most of the scripts including Ńdébé is that they are dealing with the tone…because Yoruba, Igbo and some of these languages are tonal, every script will fail if it doesn’t find a way to deal with the tonal properties of these languages. If you are thinking of a pan-Nigerian language, then we have to make sure it is a language that can be adopted all round the country.”
There are a number of interesting illustrations. Thanks, Trevor!
Yoruba, the most widely spread language in sub-Saharan Africa
Nope. Not even close.
A less popular indigenous writing language but also integral to the history of Western and Northern Nigeria, Ajami Script
Arabic script is no more “indigenous” to Nigeria than boko, though admittedly it got there first by some margin.
Ajami laid a foundation upon which present-day Yoruba alphabets found their footing
By no stretch of the imagination.
And so on …
They mean well, and one can only applaud their objectives. But they should look at Wikipedia before writing their articles. Or even, you know, consult an actual expert.
Yeah, it’s pretty sloppy. But nice illustrations!
As I’ve said before, Wikipedia provides a nice benchmark for epic fail: if you understand the topic you’re writing about less well than Wikipedia does, you really, really shouldn’t be writing about it.
every script will fail if it doesn’t find a way to deal with the tonal properties of these languages
This is nonsense.
In the real world of actual literacy work, resistance to tone marking as standard very often comes from the very speakers who are actually trying to use the orthography in question, who, after all, tend not to need it (depending on the language: Yoruba is getting on towards the Mandarin end, but lack of tone marking in Hausa, for example, really isn’t a problem for Hausaphones at all.*)
I’ve lately noticed an efflorescence of tone marks on Yoruba personal names in Western media, incidentally. While I, personally, am all in favour of tone marking (why won’t the Russians mark stress as standard, for heaven’s sake?) this looks rather as if the issue has got hijacked for political reasons.
* It’s not marked in ajami, of course. Arabic script has enough trouble writing the vowels …
resistance to tone marking as standard very often comes from the very speakers who are actually trying to use the orthography in question
Indeed.
Yoruba is getting on towards the Mandarin end
For that matter, people writing in Pinyin tend to ignore tone marking, even in handwriting where there are no technical barriers.
I have a Yoruba Bible (not that I can read it), which has only the occasional tone mark, presumably only those truly necessary to avoid actual ambiguity for native speakers. There are not many.
Yoruba is definitely at the more-significant end when it comes to tone in West African languages, too (though by no means the record holder, admittedly. For that you need to go to some of the more weird and wonderful Grassfields Bantu languages, or the Mande languages which have reduced all the disyllables to single syllables but kept all the original tone distinctions …)
With Kusaal, I have no great difficulty even as a foreigner with supplying the correct tones from context, unless I come across a word that I just don’t know at all (at which point, unfortunately, Naden’s dictionary is no help whatsoever.) This, despite the fact that Kusaal has a fair number or lexical tonal minimal pairs, and that syntactically determined tone is also very important in the language. You can work it out from context, almost always.
Swallowed the dictionary, did you?
How does untoned Yoruba writing compare to unpointed Hebrew or Arabic writing?
Swallowed the dictionary, did you?
Would if I could …
No, the trouble is specifically that none of Tony Naden’s dictionaries mark tone; he himself has difficulty hearing tonal distinctions (and wrongly described Bisa as not having any) and I think this has led him to seriously underestimate its importance.
On the other hand, the fact the he has done significant and valuable lexicographic and translation work on Western Oti-Volta despite this rather bears out the point about tone in these languages: if you’re a native speaker, you don’t need no stinking tone marks, and if you’re not, you can still understand what people are telling you even if you’re tone deaf (as it were.)
How does untoned Yoruba writing compare to unpointed Hebrew or Arabic writing?
I think that’s actually quite a good parallel.
It does occur to me, though, that just because a native speaker can puzzle out tones from context, it doesn’t necessarily mean that it wouldn’t be easier to read texts which gave more than the minimum absolutely necessary cues regarding tone. Yoruba would (I think) lend itself to more copious tone marking than most, partly because the tonal system is quite well understood, and partly because there isn’t a huge amount of tone sandhi to confuse the issue by making words in context look different from their citation forms.
In Kusaal, an the other hand, the best analysis to date of the tonal system is by me*, and it’s by no means widely known, and tone sandhi and tone overlays are pervasive. Marking tone accurately in these circumstances is actually quite difficult, and it seems very unlikely that it would be worth the trouble for native speakers to even try. It would be like asking English speakers to routinely mark all four contrastive levels of stress.
* Not as megalomaniacal claim as it may sound. Nobody else has really tried.
Navajo is an interesting case. Linguistically, low tone is unmarked, but as all reduced vowels are forced to be low, high is the orthographically marked tone. However, people who write Navajo at all, seem to write it correctly: there is no semiliteracy that I know of.
Hausa Ajami doesn’t write tone, but Kanuri Ajami mostly does – Kanuri has no phonemic vowel length, so they repurpose it for marking high tone.
I remember Phil Jaggar telling me most of the Hausa linguists he had worked with couldn’t reliably transcribe Hausa tone themselves; for some reason (possibly a problem in the accepted analysis?) it simply doesn’t seem to be particularly intuitive.
(I should say Kanuri didn’t have contrastive vowel length – now that they’ve lost intervocalic g I suppose they technically do, but all the Ajami data I’ve looked at predates that.)
I have a Yoruba Bible (not that I can read it), which has only the occasional tone mark, presumably only those truly necessary to avoid actual ambiguity for native speakers.
This version has tone marks throughout.
“Indigenous” keeps reminding me of a song my friend composed in 17 when fooling around with a piano.
It begins with: Я сюда раньше всех убежал… “I to-here befor/earlier [than] all/everyone ran-away…”
“f you’re a native speaker, you don’t need no stinking tone marks, and if you’re not, you can still understand what people are telling you even if you’re tone deaf ”
English articles.
They are difficult for Russian readers partly because without them books are no less intelligible. ʾIʿrab.
This version has tone marks throughout
Heresy!
Hausa Ajami doesn’t write tone, but Kanuri Ajami mostly does – Kanuri has no phonemic vowel length, so they repurpose it for marking high tone
That is extremely interesting. Thanks, Lameen.
The analysis of the Hausa tone system seems to be pretty uncontroversial, and it’s a pretty simple one, too, as these things go. I suppose L1 speakers just find it hard to focus on. The same thing seems to be true of Kusaal tone, too: the accounts of it by L1 speakers are pretty perfunctory, whereas I had to pay a lot of attention to it exactly because it was so difficult to get a handle on as a foreigner. (It turns out that the syntactically determined aspects of tone shed a lot of light on other aspects of the grammar, which had not been described before because tone was simply taken for granted.)
I suppose that for both Hausa and Kusaal speakers, tone seems almost as if it were subphonemic. It seems that Yoruba may be different.
@DE, I wonder if native speakers are more friendly to vowel-quality diacritics than tone marks.
Because the graphical differnece between e and é also may play a role…
As for Russian stress, our orthography could be more phonetical, and then for stressed semi-long /a/ we would use a different glyph than for reduced ⟨a/o⟩.
Sadly, there is no convenient way to type Cyrillic stress:( For English I am using a layout where
"arendersä, but I am not willing to insert those in Cyrillic text because they will render words ungooglable. Wont work with ы.On forums/chats we often use capitals for this: корОва.
Because the graphical difference between e and é also may play a role…
My print copy of the Mooré Bible does indeed use the acute mark to distinguish /e/ from /ɛ/, and indeed generalises the convention across all ATR pairs. (It uses the circumflex to mark nasalisation, too.)
The online version has abandoned this in favour of using proper symbols, though. Still no tone marking, however.
Just repurpose a keystroke combination to mean U+301 COMBINING ACUTE ACCENT (for me it’s AltGr+’ followed by =) and then you can get ы́ or anything you want.
When I was studying ESL 40 years ago my final project was on how to teach the English articles. I started off by looking at 5 or more grammars and teaching methods and found that none of them had anything useful to say — partly I suppose because, as drasvi said above, they’re usually not necessary except for making yourself sound fluent.
The basic rule is:
1. “A dog walked into the room”. First mention of dog.
2. “The dog walked into the room”. This dog is already known, a particular dog.
3. Dogs like to walk into rooms”: generalization. (A sentence unlikely to eve be said, invented as a example).
But there are a lot of exceptions, like “The dog is a noble creature”.
My Chinese teacher had immigrated to the US from Beijing via Taiwan decades before she taught me and had been a successful businesswpoman here for years, but she had no use for the unnecessary parts of English, such as tenses and articles.
In re Hausa linguists finding it difficult to mark tone correctly in their own language:
I wonder if a contributing factor is that the convention in all the grammars is to mark low tones with a grave but to leave high tones unmarked? I presume that there are good reasons for that rather odd convention (the linguists responsible being far from stupid) but I don’t recall ever having seen a discussion of them.
It may be nothing more profound than that Hausa has more high tones than low in a typical utterance, but it doesn’t seem to be by a huge margin. Maybe there are more subtle arguments, to do with markedness, in which case, you might have expected that to chime with speaker intuitions rather than conflict with them, though.
Regardless of the answer, though, it’s hardly surprising that regular tone marking hasn’t taken off among literate Hausa-speakers in general if even Hausa linguists find it difficult.
they’re usually not necessary except for making yourself sound fluent.
There’s more to it than that. Redundancy (superfluity) is critical to intelligibility, and eases understanding. This is because communication channels are often noisy and/or lossy.
For example, a non-native speaker of your language doesn’t better understand what you’re saying when you talk baby talk (no tenses or articles) and pump up the volume. That is, when you leave out what you imagine to be the “unnecessary” bits and decrease the signal-to-noise ratio.
Also, the transmission time occupied by “unnecessary bits” gives your interlocutor time to think.
it’s hardly surprising that regular tone marking hasn’t taken off among literate Hausa-speakers in general if even Hausa linguists find it difficult.
It is equally unsurprising that the use of IPA hasn’t taken off among literate speakers in general though linguists find it easy – in fact revel in it (sich suhlen darin, as we say in the Old Country).
RIL: sich sielen is also a thing.
I’m not convinced that simply leaving out articles and tenses etc in English really does produce anything at all comparable to Mandarin. It’s not that Mandarin never expresses things like givenness and time reference at all: it just does it all very differently from English. If you strip all that out from English without compensating for it in other ways somewhere you just cause a loss of clarity without any kind of gain at all.
Stu’s point about redundancy is also very much to the point. You say “Redundancy!”; I say “Sorry, I didn’t quite catch that. Could you say it again?” Mandarin has (no doubt) its own redundancies. The “junk DNA” of language, without which it Doesn’t Bloody Work, even though it looks as if it ought to.
Has any linguist every addressed the junk DNA of language ? I have the impression that linguists are generally in thrall to one-to-one representational theories. LC (linguistic correctness) appears to mandate that everything has a function or meaning. The idea does not come up that there are non-functional components which have a function – though not a one-to-one or representational function, or a meaning you can grasp.
Maybe we are subservient to these components. I see a title: The Selfish Dust Bunnies of Language.
It may just be me, but when I did Greek at school, I was grateful that (Attic) Greek, unlike Latin, has a definite article. A great practical help in finding your way around complicated sentences, even though you could mostly leave them all out and still have a more-or-less comprehensible text.
I bet Russians are not grateful for it. το ον is a helpful reification for an English speaker. Ens, say what ?
It may just be me, but when I did Greek at school, I was grateful that (Attic) Greek, unlike Latin, has a definite article.
Did they teach you anything useful about the particles? When I studied Greek, they were treated as pretty much an annoying cluster of dust bunnies (to use Stu’s excellent terminology).
Did they teach you anything useful about the particles?
Yes they did, although you got a definite feeling that they were outside their comfort zone. The particles don’t fit the neat categories of traditional grammar very well. I think there was a sort of underlying hope that you might learn their “meanings” by osmosis.
Incidentally, the Kusaal particle naan is used in ways eerily very reminiscent of the Greek ἄν, an observation which (unfortunately) seemed unlikely to be helpful to really any plausible potential readers of my grammar …
It was only learning Russian that taught me how to think about Greek particles. Actually using such things is a great help.
Definitely. I remember that the Big Reason I wanted to learn German better was in order to understand how those “particles” worked, especially when they are strung together: doch, mal, schon, eben, gerade, halt, wohl, ja, gar …
It’s all very well to be told:
# Partikeln sind Signalwörter, die eine gesprochene Sprache lebendiger machen. Sie wecken beim Zuhörenden ein Interesse oder machen ihn neugierig. Ein gesprochener Text kann mit Partikeln positiv oder negativ bewertet werden. Partikeln sind unveränderlich, nicht deklinierbar und können nicht erfragt werden.#
But how does that work in practice ? It’s like learning what A7, G7 etc are but still being unable to accompany Ella in Melancholy Baby.
It was only learning Russian that taught me how to think about Greek particles.
Learning German taught me how to think about Latin cases. Think ? No, feel.
This week I happened across the expression “accusative of respect” for a Latin “construction”. Only the examples shone a light on what was meant, for instance hanc ego nondum etiam sensus deperditus omnes … .
I see now that not knowing this kind of thing (along with hundreds of others) has hampered me in the past. But “accusative of respect” is just stoopid, I have already forgotten it. Example’s the king, wherewith to catch the nature of the thing.
Learning Kusaal taught me how to think about the English tense/aspect system.
German speakers can indeed hold their heads up high among the discourse-particle users of the World.
I think that’s a coincidence: people have noticed that the tone marks exist, and they’ve at last noticed they can actually reproduce them on their computers even though they’re not on every keyboard.
Our Serbian family Bible* does the same, marking у о̀ко̂ in at least one place to show it means “in the camp” and not “in the eye”.
(…The circumflex must be for “posttonic length”. Tones are only distinguished in stressed syllables, but two of the tones are long, and when the stress moved one syllable closer to the beginning of every word, the two long tones left “posttonic length” behind.)
* The original translation from Church Slavonic – which had apparently been read in a rather Russian pronunciation, leading to phenomena like Кајин and Авељ that my dad finds funny.
Most of the time it really doesn’t, but that’s different from never.
The second syllables most words now have come to mind. Their point is to save the language from catastrophic homonymy, but they aren’t necessary as often as they’re used.
Most of the time it really doesn’t
What I had in mind was that English definiteness is intimately bound up with other discourse-organising features, like topicalisation, focus and anaphora. You’d need to have a look at fairly large-scale ways Mandarin achieves textual cohesion to see how far it was doing the same thing as English by different methods, and how far it just wasn’t bothering with it at all.
What’s your experience with how Latin does it? 🙂
It is possible that articles facilitate reading. But I already was a very fluent reader before and I did not feel that. In Russian everything is the same, just without articles, but with topic-comment structure (overlaps with definiteness in that topics are usually definite), and there is translationese (SVO, SVO, passive, SVO, what topics?).
“I did not feel” does not mean, of course, that they do not help native speakers. “Help” in turn, does not mean that every instance helps: it is possible that for article users it is easier to generalize them for contexts where they don’t. Maintaining a complete system can be cheaper. It does not mean that the system is optimal either, and it is quite possible that for a certain set of set phrases we could change their values.
I felt the need to use them when I began to write on forums (not oral communication, but not the same as composing a letter either…) Based on this I think that the main motivation for articles is discoursive.
“that English definiteness is intimately bound up with other discourse-organising features,”
@DE, do you have some particular text (like a chapter in a grammar book…) in mind?
@drasvi, de
Leaving out articles in English gives an impression of clipped speech, e.g., as employed in the military or by headline writers. Because of the lack of and/or redundant forms for declensional and conjugational endings, this often leads to indigestible or even ambiguous phrases, i.e., the headline “Foot heads arms body!”, which the reader needs to parse as [Michael] Foot heads [i.e, is appointed head of] [the/an] arms body [i.e., whatever kind of committee or advisory board related to weaponry this was]”. In russian this could read Foot [Nominative, clearly foreign name] heads [Verb, 3ps] arms [adjective, accusative] body [accusative]. (This would be an equal number of words, but no ambiguity or need for much thought). There would even be another 4-word Russian equivalent corresponding to Foot head[noun, Nominative] arms [adjective, genitive] body [genitive]–this might even be a more typical syntax with again, no ambiguity or need for thought
do you have some particular text (like a chapter in a grammar book…) in mind?
No particular text, no; but it’s not a very controversial statement.
The immediate reason I thought of it was – inevitably – to do with Kusaal grammar, in which the basic VP-focus particle, nɛ, is potentially ambiguous, so you need to pay attention to things like definiteness to interpret it correctly in context. But, as I say, it’s not a very ground-breaking point.
Your example, that topics are usually definite, is exactly the sort of thing I meant, though. (This in fact interacts with DM’s question, above: Latin word order is greatly affected by topicalisation.)
If only Kusaal were taught universally, beginning in the early grades, so much muddled thinking could be avoided!
So true. It is the preeminently logical language. French, nothing.
(Actually, I really do think there would be a genuine benefit in teaching pretty much any sufficiently exotic language to SAE-speakers from an early age, so that they could internalise the vital message that It Ain’t Necessarily So.)
Absolutely.
I ignored my English lessons and tried to learn Old Irish.
The most boring subject in school was Russian, and that’s despite my huge interest in my language and an excellent teacher (I befriended her).
I am not sure that it would be better if they taught (desecrated…) Old Irish.
(the did not harm English though: English teachers here are not very demanding, and especially they do not want to have MORE students who are not interested).
Much as I love Old Irish, I think it would be best to use a language that one can actually speak.
@LH, I think I always loved manuscripts and only later became curious about people:(
Old Irish texts did affect me in that those people spoke/thought/experessed thoughts differently. It helps with modern people from different cultures… and also I am less afraid of texts that I do not understand, be that Chinese or molecular biology.
I do not mean that I would not be happy to have English-speaking kids in our class.
I just mean there was a moment when I thought: “hey, for all my life languages for me were texts. I am tired, and now I want to see it with a speaker’s eyes.”
I am just a bookworm.
Me too. (Obviously.)
https://en.wikipedia.org/wiki/Yoruba_people : Most Yoruba people speak the Yoruba language, which is the Niger-Congo language with the largest number of native or L1 speakers.[33]
https://en.wikipedia.org/wiki/Yoruba_language : …several other millions of speakers outside Nigeria, making it the most widely spoken African language outside of the continent.
It is not clear where their data comes from though.
There are nations that are thought to be less verbose.
E.g. the Russian joke about Finns/Estonians (a father and two sons in a cart, one of brothers says “fool”, a few hours later the other responds “fool yourself” and then a few hour later the father says “don’t fight, hot[-tempered] Finnish [Estonian] boys/guys”). In Russian “hot” about a person speaks about her temper, not looks. The joke is silly but those “hot Finnish boys/guys” or hot Estonian boys/guys are frequently quoted, e.g. when humorously urging someone to stop a jokular argument.
Compare to this Moroccans who would say to me: “no, we have lots of vowels, haven’t you heard our songs. Listen: fhskhkskhsdhfdksfh! Hear? Vowels! But I said it slowly and we just speak really fast, so you do not hear”.
So are there most vigorously/actively/passionately spoken languages (as opposed to “widely”)?
I think there can be little doubt but that there are more L1 Hausa speakers than L1 Yoruba speakers*; the original article just says “African language”, not “Niger-Congo language”, and does not specify L1. When it comes to L2 speakers, Yoruba is nowhere behind Hausa and Swahili, of course. It’s hard to know how many L1 Swahili speakers there are, and indeed whether the L1/L2 dichotomy is really valid in such polyglot communities, where L1 competence in several languages is common. Certainly most L1 Swahili speakers are not ethnic Swahili nowadays.
It’s true enough that any language spoken in Nigeria is going to be a strong contender for “most populous”, though, and it’s certainly conceivable that Yoruba really is the most widely spoken Niger-Congo language as an L1, though I share your scepticism regarding the statistics (especially having actually lived in Nigeria.)
* And if there weren’t already, there soon would be: Hausa is a Sprachenfresser, and is the very reverse of endangered.
WP gives 43 for Yoruba 50 for Hausa. There is also Arabic, I do not know where their “Sub-Saharan” begins…
There is a table in Languages_of_Africa where Yoruba is 28 millions for 2007… but Berber in that table is 16 millions…
“Berber” is not all one language, though. (Another reason for treating these sort of numbers with considerable caution. On the other hand, Xhosa and Zulu, for example, are the same language, at least as far as respectable linguists are concerned, so they should not be counted separately …)
Amharic might be another contender for “Third* Most Widely Spoken L1 in Africa”, come to think of it. The margin for error in some of these figures is pretty large. Without any doubt it’s the second most widely spoken Semitic language in the world**, and of course the third most widely spoken AA language.
* After Arabic and Hausa.
** Unless Arabic cheats by competing as several different languages. There are a lot of Egyptians …
What’s the current story on the transition from article-less Latin to the ubiquitously article-full modern Romance languages? Was it a chance innovation in late-ish but pre-split Vulgar Latin that then just got retained in the daughter languages, or did it arise in parallel post-split (maybe with some areal effects?) because of a widely shared Felt Need for articles, especially to offset the decreased availability of other syntactic strategies that had relied on the former case system?
All of the Western Oti-Volta languages have definite articles except Boulba, but they are of four different origins, and can’t be reconstructed to Proto-WOV.
Romance articles are of at least two different origins, come to that, even leaving aside the Easterners’ weird positioning habits.
Germanic, likewise. But with Norsemen. Everything’s better with Norsemen.
WP, Al-
“Etymology
The etymology of al- is the study of how it developed and how it changed over time.”
they should have stopped here….
Romance/Vulgar Latin article: some time ago I read an article that examined the Peregrinatio Aetheriae ad loca sancta, a (probably) 5th century text thought to be close to spoken Latin of the time; the question was if the frequent use of ille/illa/illud in that text somehow prefigures the later Romance definite article. The conclusion was it didn’t — the frequency of ille was a stylistic feature, the meaning was the same as in literary Latin. So it seems there was no Vulgar Latin/Common Romance definite article in the 5th century. And I think there is no article in the Straßburg oaths, the oldest surviving French text, although that may simply be because it is an extremely literal translation from Latin, ignoring syntactical differences (pro dei amore > pro deo amur).
Greek influence, as also seen in the “have”-perfect (reportedly calqued on a Greek construction with an aorist participle) and in words like French lisse (“smooth”).
(…at least in the west, see below.)
It certainly helped that Latin had a surfeit of demonstrative pronouns.
In the choice of demonstrative pronoun, but probably not in the choice of the concept in the first place.
That’s an areal feature, so perhaps Albanian has had its postposed definite article since time immemorial and Greek has nothing to do with the one in Romanian. Who knows. At least it’s documented that the Norsemen developed theirs independently.
Greek influence
What’s the source for this? It doesn’t seem inherently very plausible, since surely Greek influence was confined to the upper crust, who are a more likely source of loan words than whole new grammatical features.
That could still be where the article later came from. I’m thinking of the conspicuous “overuse” of demonstrative pronouns in Czech – they’re not definite articles, but you can see those from there.
That’s particularly likely because definite articles are there, at every occasion, in the German version of the oaths, and (like the “have”-perfect) they’re thought to have diffused from Romance.
Yeah it seems questionable that posh Silver-Ageish Latin writers kow-towing to Greek culture would not have borrowed the article concept (out of respect for their own Latin heritage) but that illiterate medieval peasants whose sorta-kinda-literate-in-Latin parish priests and even bishops were generally ignorant of Greek would have. The Ven. Bede (admittedly outside the Romance-Vernacular zone) could reportedly read Greek, but that was a noteworthy fact about him because it was so unusual in Western Europe by then even among what passed for the intelligentsia. (According to a moment’s googling there is no consensus as to whether Alcuin could read Greek, and one pious source that confidently claims that he could also claims he possessed literacy in Hebrew, which seems … even more of a stretch.)
surely Greek influence was confined to the upper crust
No, there’s a whole set of yobbo Greek loans in Vulgar Latin as well. It wasn’t just the nobs.
And Trimalchio is no aristocrat …
I’m not sure that it’s actually necessary to assume Greek influence though. Development of definite articles is not rare. (I’m prepared to go out on a limb here and say that Greek influence was probably not a major feature in the development of the Western Oti-Volta articles. Nor in Arabic or Hebrew.)
It does seem to be very prone to spreading across Sprachbunds, though. WOV, Buli-Konni, Gurmanche, Mbelime … all have definite articles, and mostly they do it quite differently, but the idea is the same. It’s a contagious concept (like having grammatical gender …); if your neighbour’s language has it, you want it too. I’m only surprised that Peter the Great didn’t see the need for it in Russian.
As usual I can’t remember sources, but it seems to be consensus that the Empire was a pretty mixed place linguistically, and that late western Vulgar Latin contained, at least, a heap of Greek words. There’s probably more controversy about grammatical influence, but I wouldn’t be surprised.
I don’t think we know when the old Greek colonies on the French and Spanish coasts stopped speaking Greek.
Don’t forget that a lot of Southern Italy must have switched from Greek to Latin / Romance at some point, and that there must have been a lot of slaves, freedmen, and immigrants from the Greek-speaking East all over the Western empire. With Greek being a prestige language, they probably had less inhibitions transferring features to their Latin than speakers of “barbaric” languages like, say, Gaulish or Iberian.
EDIT: partially ninja’d by DM.
“At least it’s documented that the Norsemen developed theirs independently.”
Postposing vs. preposing can be areal.
Cp. Ernst Pulgram, Italic, Latin, Italian, p. 221 (commenting on an extract from Petronius’ Satyricon containing Greek loanwords alien to literary Latin):
I mean, the areal can technically include both Albanians and Norsemen.
Because people who live in between postpose…
Surely some Chomskyan has got round to explaining that, really, all articles follow their nouns*, and it’s only our predilection for butterfly collecting that prevents us from seeing the Fundamental Truth?
* Or precede. No matter …
Regarding the concept of definite articles:
I think it’s clear that a linguistic concept/category can be inherited or borrowed quite separately from its actual expression in a particular language: they transcend their fleshly incarnations. Diachronic examples are quite common: languages change how a tense or aspect is expressed, for example, but the oppositions expressed in the basic system are preserved.
Polotsky’s deeply ingenious study of Egyptian leveraged this very thing by proposing that Middle Egyptian had its own version of what Coptic does with its “first” and “second” tenses; later Egyptian had kept an old distinction but expressed it periphrastically (and then it all got fused together afresh in Coptic.)
https://en.wikipedia.org/wiki/Standard_Theory_(Egyptology)
(Unfortunately this actual article makes a pig’s ear of explaining it. What it unhelpfully calls “emphatic* sentences” are in fact sentences with focus on an adjunct; it was Polotsky himself who discovered that this is what Coptic “second tenses” do.)
* “Emphatic” in a syntactic context, is handwaving, and essentially means “I don’t really understand what this construction actually signifies.” It’s a dead giveaway.
Good points all about Greek influence; as usual, I was shooting from the hip.
Lots of non-posh Greek influence in Italy proper (as well as posh, at least as long as dealing-with-the-Byzantines remained a thing that the political class did), but maybe not so much in other parts of the Romance Zone, which becomes rather important if the consensus is that articles developed in parallel after the breakup of late Vulgar Latin?
Vulgar Latin terms of Greek origin underlie e.g. French jambe, coup, corde, parler, pierre, épée …
https://en.wikipedia.org/wiki/Lexical_changes_from_Classical_Latin_to_Proto-Romance
(That’s before you get on to the specifically Christian stuff like évêque, église, blâme.)
In derivation, you have the verb-deriving suffix -iser.
Do we think “pierre” as a common noun for “stone” entered Old French before “Pierre” as a specifically Christian proper name did? Or was it a package deal? (Heck, did “petra,” said to be a “late borrowing” from Greek, enter Latin before Latin translations of Mt. 16:18 that wanted to make the pun explicit did?)
A few decades ago Greeks were a thing both in Marseille and North Africa (maybe they still are).
For an empire whose large (and important) part is the hellenistic world the degree of Greek presence in the west can vary widely. We have no idea how many people speak Persian in Bukhara right now (or in Moscow to that matter). How can confidently say anything about any Roman town?
My all-time-favourite I-can’t-believe-it’s-actually-Greek word in English is fleam:
https://en.wiktionary.org/wiki/fleam
I blame Galen.
How could I forget la pierre!
Perhaps not, but that would be early enough.
The people who live in between don’t use articles today, and didn’t in earlier times either, although there’s a related phenomenon in the region: both Germanic and Balto-Slavic marked definiteness on adjectives before they had any hint of articles, and they did so with non-cognate but postposed means (n-stems in Germanic, a suffix derived from the PIE relative pronoun in Balto-Slavic).
Italian does actually have an article-related commonality with Greek that the rest of Romance (not to mention Germanic) lacks: the use of definite articles together with possessive pronouns (il mio …, o … mou).
the use of definite articles together with possessive pronouns
Kusaal does this, but specifically because a possessor doesn’t make the possessum definite unless the possessor itself actually has the article. So when you introduce “my child” into the discussion first, you just say m biig, but when you refer to her again, you say m biig la. I think that this probably means that “possessors” are really more like embedded datives than SAE genitives: “(a/the) to-me child.” (Pronoun possessors in Kusaal head their own NPs: they’re not dependent pronouns, like “my” in English.)
Italian does actually have an article-related commonality with Greek that the rest of Romance…lacks
Articles plus possessive pronouns are also used in Portuguese, Catalan and Romanian – so Spanish and French are arguably the odd ones.
French has le mien…
Definitely. Petra occurs in the works of Seneca and Pliny the Elder (and others).
La Cosa Nostra (that term may have been invented by American mobsters; Joe Bonanno — who, of course, disliked it — claimed first hearing it in the late 1920s)
Huh, I actually knew that about Portuguese and then forgot.
Given the morphology involved, that’s more like der meinige than meiner, I think.
Lots of non-posh Greek influence in Italy proper (as well as posh, at least as long as dealing-with-the-Byzantines remained a thing that the political class did), but maybe not so much in other parts of the Romance Zone, which becomes rather important if the consensus is that articles developed in parallel after the breakup of late Vulgar Latin?
Italy was the most densely settled part of the Western Empire, and, more important, the part from where cultural influence radiated to the rest. That would as well work for a tendency to use demonstrative pronouns (but different ones) in an article-like manner.
“The people who live in between don’t use articles today, and didn’t in earlier times either, although there’s a related phenomenon in the region:”
Yes. I just meant, the positioning habits can be an arealism.
For modern Slavic: a usual feature of Russian dialect speech is the particle -to. It recembles articles etymologically and it is adjacent in function (different in different dialects). By “adjacent” I mean: I think if dialects began developing definiteness, they would repurpose -to for this.
I do not know how come that it is abundantly present (north especially*) in recorded dialect speech, in imitated rural speech but limited in literary Russian, old literary Russian and in any kind of urban speech I know (namely: modern literary Russian speakers and vernacular in 19th century literature).
* quick googling:
роботу-ту тежолу-то не можот робить-та, а сидет-то можот, гледеть-то, ходить-то кругом-то.
What I meant by this is that one’s idea of a language’s inventory without dialect data can be incomplete:(
A few decades ago Greeks were a thing both in Marseille and North Africa (maybe they still are).
Dunno about the rest of North Africa, but there are plenty of Gritli families in eastern Libya – Muslim Cretans who arrived as refugees in the late 19th century. I’ve heard claims that some of them still speak Greek.
It doesn’t really belong here, by the way, but I think some Hatters will be interested in this – a generative linguist lamenting that no one wants to do (what he considers) syntax any more, or gets jobs in it, and blaming it on Minimalism. The wheel turns…
No syntax for you
Who would have thought that the doctrine that real syntax has nothing to do with meaning might lead to a dead end? I’m sure nobody could have seen that coming.
tl;dr — With (very) few exceptions, generative syntax in academia doesn’t exist, not even among those departments that profess to be generative or generative-friendly.
If only my grad-school self could have caught a glimpse into the future! That and the fall of the Soviet Union would have competed for prospective joy.
I do occasionaly see references to the Greek community in Tunisia in unexpected contexts. Fishery, in particular sponge, the “unexpected context” is Cousteau or other texts about the sea. There were also Tunisian officials of Greek origin (e.g. Mustapha Khaznadar and his brother) but it is a different story: they were slaves.
I do not know details and do not know if it extends into Algeria..
Quick googling brought: The Greeks in Tunisia during the 19th century. Social adaptability,
commercial enterprise and political flexibility. (link) with a line: “The fall of the Qaramanli dynasty in Tripoli of Libya in 1835 will lead many Greeks to move west and to settle in Tunis. “
prospective joy.
я проснулся в шесть (or пять) часов
с ощущеньем счастья
нет резинки от трусов
и советской власти!
The version from my Russian language teacher (i told, my teacher was excellent, but lessons were still boring). The paradoxality of it amused her.
A rude continuation from my classmate is вот она! вот она! На хую намотана!
Khaznadar – funnily similar to archaic Russian kaznačéj “treasurer” < kazná "treasury" < Turkic < Iranic?
я проснулся в шесть (or пять) часов
с ощущеньем счастья
нет резинки от трусов
и советской власти!
I love it!
J.W Brewer, David Eddyshaw, David Marjanović, and any/all other interested Hatters: on the subject of Greek influence upon Latin with particular reference to the genesis of definite articles in Romance, I refer you to this LH post and thread from a long time ago, which is the very first one where I commented, I believe (March 17, 11:21):
https://languagehat.com/kalendae/
I stand by everything I wrote back then, and believe my arguments (My March 18, 9:47 and March 19, 1:16 comments) remain valid.
Thanks, a very interesting thread to revisit. Here is your first comment.
I wonder why constructions like Ἰωάννης ὁ βαπτιστὴς are productive in Romance and English (Peter the Great and so on).
Not only English, it’s Johannes der Täufer in German as well. It’s just an apposition identifying that specific John out of all other people / saints with that name.
It is widespread in the West, both with nouns (the Baptist) and adjecties. But you do not say “table the big” or even “Cairo the greater”. It does recemble Greek (which was not popular…) and it also recembles a number of things like Arabic…
Well, I say “table the big” when I want to distinguish it from other tables in the referential environment.
I am defeated. But my curiousity is not.
In Kusaal the second element of an apposition has to be either a personal name or a self-standing relative clause: you can say (as it were) “the Baptist John” or “John, he who is the Baptist”, but not “John the Baptist.” (You actually can’t say “John who is the Baptist” either, because personal names can’t do the necessary compounding with a dependent pronoun acting as a relative, so apposition is the only possibility.)
Who would have thought that the doctrine that real syntax has nothing to do with meaning might lead to a dead end? I’m sure nobody could have seen that coming.
I could have, but nobody ever asks me. For as long as I can remember, I have taken the word “real” used in that way (also “really”, “essentially”, “fundamentally”) as a sign that something is fishy.
Only much later was I able to identify it as a sure sign, and also name the genus of fish, thanks to a rigid regimen of Advanced Thinking.
I am defeated.
Don’t be. That’s Stuish, not English.
a genuine benefit in teaching pretty much any sufficiently exotic language to SAE-speakers from an early age
just chiming in to say YES. and especially for english cradle-speakers, since we tend to be a monoglottic lot (in the u.s., by deliberate state policy, among other factors).
i’m eternally grateful for my one early-ish (high-school) year of kiswahili, though i never reached anything like a functional level with the language.
Well, I thought drasvi was dealing me a bit of deflective irony. I couldn’t tell for sure, so I was defeated too.
John the Smith (and Jan te Winkel)
—
Stu indicaded that it’s individual, and I do not know how common it is in English in general. It happens that I’m tempted to do so humorously. It was a counter-exmaple to my claim as I formulated it, but whatever the level of commonality of the construction in English is, it does not cancel the question (how did it get in there).
I do not know how common it is in English in general.
It is nonexistent in English in general, except in the kind of titles you mention (e.g., John the Baptist).
I’d say it’s grammatical (but not pragmatically neutral) when used with named entities: America the beautiful, Fido the brave. Some adjectives don’t work as well, though: *John the intelligent, *Charlie the lazy, *Florida the sunny. Maybe someone wrote a paper about it.
“Truth social” might not work very well either. It’s hard to tell, since I have no idea what it could mean.
Eggs Benedict !
What the Fuck…
They’re productive exclusively for official name extensions like your examples. Keep in mind that Germanic adjectives, and certain Romance adjectives like “great”, can’t be postposed unless they’re nominalized, in which case they need an article anyway.
“TRUTH Social” isn’t supposed to mean anything, it’s supposed to be SCREAMED so loud you can’t think.
I don’t think they need to be official at all.
Giles the geologist; Fred the pharmacist; Helen the hairdresser … Winnie the Pooh …
To say nothing of Welsh usage. “Oh, it’s Evans the Spy you want …”
Ethel the Aardvark …
The construction does lend itself (although not inevitably) to an archaic-poesy sort of feel, e.g.:
“My shadow follows me,
Wherever I should chance to go”,
John the Gun did say.
“If you should chance to meet me,
As I wander to and fro,
Sad would be your day.”
–written c. 1970 by Sandy Denny (1947-78)
It works with all kind of anarthrous names: Mars the planet, Paris the capital…
To comrade Netto, the steamboat and the man.
But the kettle of fish was sly(ght)ly changed. The discussion began with Name the Adjective, Name the Noun is not quite the same.
I’d say those are ordinary additions to a sentence, flanked by commas: “Mars, the planet (as opposed to the deep-fried bar), is […]”
@DM: my point is that “Peter the Great” or “John the Baptist” are basically the same thing, only they have become a constant epithet.
I’d say those are ordinary additions to a sentence, flanked by commas
Definitely not. In fact, I would contrast, both in pronunciation and meaning
“I’ve just seen Polycarp the postman.” *
with
“I’ve just seen Polycarp, the postman.”
The first would be something I might say to someone who knows Postal P just as well as I do, and I feel needs no prompting as to who he might be at all; the second, to someone who either needs reminding what P does for a living, or is liable to confuse him with his cousin Polycarp the politician.
The first has greater stress on POSTman than on POLycarp; the second, equal stress on the two, and the intonation contours are also different, only the second having the “afterthought” intonation, and with an intonation fall on “Polycarp” in the second sentence, but not the first.
* Substitute “Methodius the mailman”, for all you Americans.
As D.O. rightly points out, though, “Noun the Adjective” is indeed different to “Noun the Noun.”
“Noun the Adjective” really is hifalutin and nonproductive. I think …
The is probably related (as DM implies) to the fact that “the Adjective” (without a noun) is itself very limited as a possibility in modern non-hifalutive (mostly to plurals of nationality, like “the French”, I think. Didn’t we discuss this before one time?)
I have to say that e.g. “Derek the Short” would immediately suggest to me, not any short Derek of my acquaintance, but some historical figure that I had unaccountably not heard of before. It would really be the only way to construe it. That, or a bit of Wodehousian whimsy.
We talked a bit about postposed adjective in English (although generally without “the”) previously here.
I was thinking, rather, of noun phrases of the form “the Adjective.” Possibly I merely dreamt it …
https://www.youtube.com/watch?v=250MMq0fTrU
(I hear no commas here …)
I’ve had a look at the list of Russian monarchs in Finnish*.
There are:
Oleg Viisas (‘wise’, Oleg the Prophet)
Iivana III (Iivana Suuri)(Ivan III the Great)
Iivana IV “Julma” (Ivan IV the Terrible)
Ladislaus IV Puolalainen (the Pole)
Aleksei I (this doesn’t have Hiljaisin appended, as a translation of clementissimus**)
Pietari Suuri (the Great)
Katariina II (Katariina Suuri)
*Normally, modifiers precede nouns in Finnish.
**There are two errors in two words on this page:
“tres grasieux”
The Adventurer (UK title: Michael The Finn; original title Mikael Karvajalka) is a novel by Finnish author Mika Waltari, published in 1948. It is a fictional tale of a young Finnish man, Mikael Karvajalka (Hairy-foot), set in 16th century Europe.
https://en.wikipedia.org/wiki/The_Adventurer_(novel)
“Oh, it’s Evans the Spy you want …”
I regret to state that this Evans is a M.O.T:
But this one can probably be attributed to the Cymry:
@David Eddyshaw: Cab Calloway’s singing doesn’t seem to have changed much over the decades, but sadly, by the time of The Blues Brothers, he looked quite a bit less spry on stage. Compare this version of “Minnie the Moocher” from a quarter century earlier; his movements are so fluid, you can see how natural it was to adapt him into an old-fashioned, rubber-hose-limbed cartoon character.
That cartoon is wonderfully surreal.
In the astonishing Homeric catalog of English personalities, places, and folkways included in Viv Albertine’s wonderful late-career “Still England,” she adopts the NAME THE ADJECTIVE form for artistic effect, describing certain of her Seventies musical peers as “Sid the Vicious” and “John the Rotten” before getting on to “Alfred the Great” in the next line.
Hi there!
I’m the author of the “No syntax for you” post that people here linked to and commented on. I’d like to clarify a few things that I’m afraid might be getting misconstrued here (though if they’re not, then all the better):
1. Nobody, certainly not me, thinks that “syntax has nothing to do with meaning.” Given that (literal) meaning is computed on the basis of syntactic structure, syntax and meaning are intimately connected.
2. What I am guessing you’re referring to is the hypothesis that, while syntax informs meaning, meaning does not inform syntax. This is not a “doctrine” as you call it, but a hypothesis. It’s important to note that this is a hypothesis about linguistic competence, not about use of language. Obviously, when one engages in language use, one chooses the utterance one wants to utter (or one’s interlocuter chooses their utterance) in such a way that its syntax will give rise to the desired (literal) meaning. It is also not a hypothesis about language diachrony.
3. The division between competence and performance is an *abstraction*. Like, say, Newton assuming that all objects are points in a 3-dimensional frictionless vacuum. An abstraction is neither “right” nor “wrong” – instead, it is a research heuristic which you can choose to adopt or not choose to adopt, as you see fit.
4. BUT, it is important to note that the hypothesis in (2) is only contentful in the context of the abstraction in (3). That is, if you reject the abstraction in (3) (as is anyone’s right to do, of course!), then the hypothesis in (2) is irrelevant.
5. Okay, so with all that out of the way, we can ask: has this led to a “dead end”? Depends in what sense you mean… theoretically and scientifically, not at all! There is a ton of empirical support for (2) (I linked to some of it from the very blog post mentioned above, but there’s a lot more), and there’s plenty of questions that fall within its scope that still require investigation, in many different languages and language families. (Personally, I spent most of my academic career working on Basque and on the K’ichean languages of the Mayan family.)
6. What *is* unfortunately happening is that funding sources for this kind of work has dried up, and, a decade or two later, departments have started discriminating against this kind of work in their hiring practices.
7. Now, if you agree that funding for this kind of work *should* dry up, then surely you have no problem with (6), either, and it’s every individual’s right to make up their own mind. But it’s not because anything has reached a “dead end” in any sense except dollars… And if you think the people allocating the dollars do so with intrinsic wisdom, well, let’s just say you’re setting yourself up for some disappointment.
8. So, there’s plenty more to do in this research area, and plenty of people interested in doing it. Does that mean any of the people commenting here have to be interested in it? Of course not. But I assure you, no dead ends were encountered, at least not in my experience. This is perhaps where I should also point out that I’m resigning from what was a tenured position as an Associate Professor at the University of Maryland. So even from the parochial perspective of my own personal career, this was hardly a “dead end.”
So feel free to be interested in what interests you, and not interested in what doesn’t. (And of course you don’t need me to tell you that.) But I hope to have corrected some potential misconceptions about what my post was & wasn’t saying.
Sincerely,
– Omer
“Sid the Vicious”
Aha! I found that for some reason *Mary Egyptian (Mary of Egypt, Mary the Egyptian) sounds particularly wrong to me, but was not sure if it is because of the pattern Noun Adjective itself or because it is nisba* or because of n./adj. ambiguity.
So I was trying all other sorts of Noun Adjective names. Sid Vicious is great.
—
Is there an English synonym?
Is there an English synonym?
In English-language Egyptology, the form you see is nisbe, presumably because Alan Gardiner liked it.
Nisbe, of course, is English; same as sandhi and Sprachbund. All of them good Anglo-Saxon monosyllables.
Talking of Saints Mary, for some reason, the 1996 Kusaal Bible Meeri one yi Magdala “Mary, she who came from Magdala” has got throughout transformed in 2016 into Mɛɛri onɛ an Magdalen “Mary, she who is Magdalene” or just Mɛɛri Magdalen. I have some vague memory that there is some question about whether her epithet actually does mean “from Magdala”, which presumably explains it.
I regret to state that this Evans is a M.O.T
Deep cover. But our peoples are of course in any case close kindred. We VSO peoples share much, though I believe that our modern Israeli cousins have now deviated from the One True Word Order. A passing aberration, it is to be hoped.
Stephen Goranson, last month:
Uzbek is crazy.
When I type in Google nisb… it suggests nisbat va proporsiya. All right, lets google for it. The first video is Nisbat va Proporsiya. Matematika 6-sinf. 17-dars
sinf is a Russian calque and Arabic borrowing at once: Russian class means a year in school (grade), Arabic ṣinf is “class”… Given that I know some Russian, some Turkic, some Arabic and some Persian (enough to recognize va), reading this mixture…. An unique feeling.
It is like you’re Frisian and studied French and then come across a text in English.
@Hat: thanks. So much for my memory …
“Holy Michael Archangel, be with us in the day of battle.”
Mmm, I’d have some schadenfreude to note from reading something like this, but I want to hear out first his take about the more general state of linguistics / humanities / academia. Going by reports elsewhere I am not ruling out that it’s, in fact, the American university system imploding in general and taking its toll on local schools / communities of thought, whether those might deserve to implode or not. Or same with some catastrophe of slightly lesser scope, e.g. the “cultural studies is eating everything” problem that I’ve recently heard a colleague from a Respected European University complain about…
Bumping this to call attention to Omer Preminger’s comment from a couple of days ago that got lost in the spam file until he called my attention to it; he has interesting things to say about syntax, meaning, and funding.
@Omer Preminger:
Thanks for interesting comments (and general good-natured forebearance with my ignorance.)
I agree that “dead end” rather begs the question, given that your program has undoubtedly generated, and continues to generate, much work of great interest to those who accept your premises and your dichotomies (as you lucidly explain.)
Is your “hypothesis” not somewhat protected from refutation by a limitation of the domain of “syntax” to exclude obvious counterexamples, and/or by leveraging the performance/competence dichotomy to discount them as genuine counterexamples? (This is what I was driving at with “real.”) I can conceive that “syntax” in this restricted sense might yet be eminently worthy of study, as you say, but that seems to entail regarding your hypothesis itself as a kind of research heuristic rather than a hypothesis strictu senso.
What I am (almost certainly wrongly) implying is that the empirical support for your hypothesis is real enough, but that the research program is itself constructed in a way that tends to exclude the kind of evidence which would not support it.
(Respect regarding Basque and K’ichean, by the way.)
I find myself wondering about the semiotic significance of Omer Preminger having The Simpsons‘ Otto as his avatar.
Otto Preminger.
I wouldn’t say that Dr. Preminger is a complete dead ringer for Otto, but if you google up some photos of him there’s certainly a closer resemblance than any of my own old professors (in linguistics or otherwise) bore to Otto. (I feel like if the Simpsons had already been around in the mid-Eighties when I was an undergrad we would almost certainly have identified some Simpsons character that Larry Horn resembled visually once you pointed it out, but I’m not going to work that one out retroactively.)
He has another post previewing some potential future posts in the “ills of academic linguistics” genre, which even non-enthusiasts regarding generative syntax might want to keep an eye out for. https://omer.lingsite.org/blogpost-the-ills-of-academic-linguistics-outlook/
It might be a modest way of letting people know his height and academic credentials:
# He is also confirmed to be 5 ft 10 inch (5’10), on the back flap of the Season 15 box set, showing said license to hold the guide.
Otto has apparently graduated from Brown University, as shown in “Lisa Gets an “A””. However, due to this information being revealed in a dream, it might be non-canon. #
Edit: this hypothesis finds no support at his homepage.
@DE: but that seems to entail regarding your hypothesis itself as a kind of research heuristic rather than a hypothesis strictu senso.
A hypothesis can always be used as a research heuristic, if one is so minded. Even when the research is only intracranial. It proceeds by trial-and-error against a background of experience. It’s the way I program, in fact.
What is the sensus strictus to which you advert ?
It was an Ultraspoonerism.
Kentucky schreit ficken, eh what ?
Like that, yes, but less über.
I suppose an even more recherche choice of image than Otto from The Simpsons would have been Thranduil the elvenking. (Listen here for the connection.)
Omer Preminger: Thanks for taking the trouble to write such a clear reply.
I have to admit, at the moment I’m finding it impossible to conceive of any interpretation of the hypothesis that “while syntax informs meaning, meaning does not inform syntax” that isn’t self-evidently false. Based on your comments, though, I realise there must be at least one; I don’t suppose you could summarise it?
(Personally I do not intend to jump on any blanket anti-generativist wagons. I’ve found some aspects of generative work quite useful and insightful – notably Cartographic Syntax – but I get the impression that such work overlaps rather considerably what you’re criticising as having crowded out syntax proper…)
Omer Preminger’s response is a model of civility, clarity, and good humor and he should be invited to continue to comment here.
I am totally objective on this because I know nothing about this issue in question and care less.
Hi again, and thanks for the kind words by various commenters here about my previous comment.
Let me reply to a couple of questions that came up in your replies.
@David Eddyshaw asks whether it is fair to say that the generative syntax research program is designed in such a way to exclude, by fiat, the kind of data that would disconfirm it. This is of course a legitimate concern: excluding data from consideration must be done on grounds that are independent from the question or hypothesis currently under investigation, otherwise the whole thing starts to resemble butterfly-collecting more than hypothesis testing. On the other hand, I find it useful to remind myself (and others) that (a) no one has a “theory of everything” from the get-go, (b) no single kind of data is privileged a priori compared to others, and (c) all data is filtered through hypotheses. Or, to paraphrase (c): we, as humans, don’t actually have unfettered access to any “facts” that are not filtered through a series of linking hypotheses. How do these three points interact with the previous one? Well, suppose I’m investigating a particular syntactic pattern, and there’s a pesky little data point that doesn’t conform to the predictions of my theory. Excluding it on the grounds that it reflects a “processing effects” extrinsic to the syntactic system is meaningless speculation, unless there is either some explicit proposal on what the processing pressure is and how it works – or, failing that, at least an independently decidable criteria for when a sentence would or wouldn’t fall under the category of utterances that are subject to these “processing effects.” (Otherwise we’re back to butterfly-collecting.) But, and this is I think important, there is nothing intrinsically problematic about excluding data if the methods of exclusion do meet these criteria, because there is literally no science that doesn’t do so. One might prefer a theory where processing effects and syntax are not separated in the first place, and that’s of course fine; but as long as one doesn’t use “processing effects” (I’m continuing to run with this example, though hopefully it’s clear that I’m trying to make a slightly broader point) as an ad hoc dustbin for inconvenient data points, I see no circularity in the endeavor. The $64,000 question is then whether generativists abide by these guidelines, or fall into the butterfly-collecting pattern of conduct. And the answer, as you might expect, is that some are better than others, and no community of scholars should be judged by its very worst practicioners. (Even when some of those practicioners are very famous! )
@Lameen asks if I can provide a concise and contentful interpretation of the hypothesis that “syntax is independent from meaning.”. I’ll give it a shot. First, observe that the primitives of syntax do not themselves correspond to meaning primitives. So, for example: in many languages “verb” (a syntactic notion) is in more or less decent alignment with “predicate” (a semantic notion), but, e.g., many Iranian languages resort almost exclusively to what we might call “light-verb constructions” (_to take a nap_ as opposed to _to nap_), having what amounts to a closed class of actual verbs. So “verb” as a syntactic category does not, in fact, align with “predicate” as a semantic category (or with any other semantic category for that matter). To cite another example, “nominative case” is dissociable from literally any semantic notion (there are nominative patient/theme arguments, for example in passive clauses; and there are accusative agents, for example in so-called “ECM” sentences, e.g., _Janice believed them to have run home_). And so on and so forth. But maybe the *structures* that syntax builds are themselves reflections of meaning? Again, it looks like surprisingly often, the answer is no. For example, a rough approximation of the semantics-pragmatics of the passive voice in English is that it’s used when the subject is topical (old information) and the by-phrase, if present, is focal (new information). But this, it turns out, is not universal – some languages lack passive entirely, while others use it for nearly any dyadic verbal construction (mostly or entirely eschewing active transitive clauses), regardless of the new/old-information status of the arguments involved.
None of this denies that there will be corners of syntax that align, probabilistically if not combinatorially, with semantic primitives. (E.g. a genitive noun phrase or grammatical possessor will often be associated with the semantic relation of possession.) But the extent to which these things are always counter-exemplified somewhere is quite striking. (To go back to the possessor example, it’s hard to say in what sense _the universe_ is a semantic “possessor” in _the universe’s inception_ without emptying the notion of any predictive semantic content.)
To repeat something I said earlier, this doesn’t mean that *the choice between structure A and B*, when for example a speaker constructs an utterance, is not driven by semantic factors. It obviously is. But that is because, when pairs of structures *do* map onto distinct meanings, speakers know which one of the distinct meanings matches their intentions, and choose accordingly. Making such a choice then ends up, in practice, as a choice between two syntactic structures. But it’s not because the structures are in and of themselves *meaning structures*; it’s because, when pumped through the engine of semantic interpretation, each structure (in this constructed hypothetical) gives rise to a different meaning.
‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑‑
On a lighter note: the choice of the Simpsons’ Otto as my avatar was made because I wear headphones a lot… It was not until afterwards that someone pointed out that, well, people might think it relates somehow to Otto Preminger!
no one has a “theory of everything” from the get-go … all data is filtered through hypotheses
With this I am in total agreement. Theory-free inquiry is an illusion: a damaging illusion, indeed, inasmuch as it entails blindness to one’s own assumptions, which therefore never get questioned.
I think the only points of (potential) difference would be in how one reacts to disconfirmatory evidence (and I take your point about not judging an entire field by its least methodologically-scrupulous practictioners; I am familiar in all too many areas with the phenomenon of allies who I wish were, ahem, not actually on my own side); and whether you have an effective program for actively seeking disconfirmatory evidence (which may involve a sort of meta-critique of the hypothesis itself, if it’s framed in such a way that it effectively discourages such efforts in favour of amassing “confirmations” – what I presume you mean by “butterfly collecting.”)
I wasn’t (necessarily) implying, in suggesting that that your “hypothesis” has more of the nature of a metaphysical research program, that it should therefore be summarily abandoned at the first false prediction: there can be a methodologically proper case for explaining away “apparent” disconfirmatory evidence in some circumstances (I’m with Imre Lakatos on such issues.)
None of this denies that there will be corners of syntax that align, probabilistically if not combinatorially, with semantic primitives
I think this is the only point that I would really insist on (though I think, myself, that the “corners” are pretty big corners.) It does seem to me that in saying this, you are limiting the domain of “syntax” to something rather less than what I myself would understand by the term, but that does not mean that this residue (or “corners”) is not worthy of study in itself. All sorts of issues suggest themselves to me regarding your specific examples, but that doesn’t undermine your general point.
BTW, I do appreciate your courtesy in the face of some provocation; I know we keep rather harping on this, but I think we have contrived to drive away some generativists from the Hattery by being a bit too partisan on these issues, which is a great pity and not conducive to learning anything. Speak on!
Not sure if this is quite what you mean, but I was thinking of the Construction Grammar people; I think that they are, ultimately, right in their conception of how language works (a conception in which meaning is absolutely central), but even the most hardcore it’s-constructions-all-the-way-down constructivist would presumably freely admit that in any given language there seem to be an awful lot of regularities across constructions of quite different meanings, and that those regularities seem a bit – unmotivated – as far as meaning goes.
BTW, I do appreciate your courtesy in the face of some provocation; I know we keep rather harping on this, but I think we have contrived to drive away some generativists from the Hattery by being a bit too partisan on these issues, which is a great pity and not conducive to learning anything. Speak on!
I am in hearty agreement, and am glad you are gracing these halls with your civil presence. (I fear I am a frothing anti-Chomskyan, but that is due to grad-school traumas that I cannot overcome. I am rather to be pitied than scorned.)
I certainly agree that both what falls inside my (parochial) definition of “syntax” and what falls outside of it are *definitely* worthy of study. And through that lens, I guess my point is that it’s getting harder and harder, from an institutional perspective, to study the part that falls inside of it.
Regarding Construction Grammar, I think it boils down, in part, to what kinds of phenomena one is interested in. For example, consider sentences involving: (i) constituent questions; (ii) relative clauses; and (iii) topicalization. These are, in one sense, three “constructions.” But it turns out that they have things in common that, when viewed from that perspective, end up looking rather coincidental. Syntactic islands are an often discussed example, but let me actually talk about something else – all three constructions license what syntacticians call “parasitic gaps”: they allow a gap inside an adjunct clause to be co-construed with an element outside that adjunct clause:
(1) Which paper did you file [without reading ___ first]?
(2) This is the paper that I filed [without reading ___ first].
(3) This paper, somebody filed [without reading ___ first].
This is noteworthy because these types of gaps, contained inside adjunct clauses, are typically illicit. For one thing, one cannot do this when the part outside the adjunct clause is a plain declarative:
(4) * I filed this paper [without reading ___ first].
(This sentence is licit on the somewhat odd interpretation that it was “without reading (in general) first” as opposed to “without reading specifically the paper referenced by _this paper_,” but that more general interpretation is not the one we’re aiming for here.)
And furthermore, construing a wh-phrase (like _which paper_) as being directly associated with this kind of gap position is also degraded:
(5) * Which paper did you voice an opinion / your opinion [without reading ___ first]?
The difference, of course, is that in (1)-(3), the clause-initial constituent is construed with another position, the object position of the verb _file_, whereas in (4)-(5), no additional co-construal of this type is going on.
But why should that matter? Why would all three “constructions” behave this way? On the generative approach, the answer amount to a paucity of means: there is only one mechanism in grammar (we call it “wh-movement” or “A’-movement”, but the names don’t matter of course) which is capable of associating a dislocated constituent with another position (as in the relation between the clause-initial noun phrases in (1)-(3) and the object position of the verb _file_), and that mechanism happens to have the property of licensing “parasitic gaps.” This, of course, furnishes a second question, which is why this would be so – why would this mechanism have this property. That is a worthy question in its own right (and there are good answers on offer; I’m partial to the one put forth in Jon Nissenbaum’s 2000 dissertation). But even independent of that, it already answers the question of why all three “constructions” in (1)-(3) would behave alike: they behave alike because there’s only one mechanism they could be based on, and that mechanism will either license parasitic gaps or it won’t, and so all three will necessarily behave alike in this respect.
On the Construction Grammar view, it’s a lot less clear why this would be. Why couldn’t we as speakers of English or any other language learn, e.g., a relative-clause construction that allows parasitic gaps, alongside a wh-question construction that doesn’t? (It’s logically possible to give an answer whereby the “constructions” are a lot more abstract in nature than “wh-questions” and “relative clauses” and the like, and instead there is a meta-construction (if you will) involving dislocated elements, and (1)-(3) are all based on this meta-construction. But at that point it’s not clear that the two accounts differ anymore.)
Lastly, at the risk of repeating myself, I do really appreciate the kind words from you all. I do my best to behave well online (and do not always succeed), and the community here deserves kudos of your own for how you reacted to my initial comment here. I’ve seen many less hospitable online spaces.
I’ve seen many less hospitable online spaces.
Oh, I think we all have.
@Omer Preminger: I would be leery of any syntactic argument that distinguished your (4) from (1–3) as operating under different mechanisms, because I find, “I filed this paper without reading first,” to be grammatical. However, it is awkward, in a way which I interpret as indicating that it is almost ungrammatical. Other native speakers might find that, in their idiolect, (4) lies on the other side of that line and is frankly ungrammatical. In fact, I think your example (4) does a very good job of showing that grammaticality judgements can be scalar, rather than strictly binary; moreover, they can be intersubjectively nonuniform.
On the other hand, I’m not sure what you are driving at with your example (5), since it appears to be malformed (missing a preposition) even without the, “without reading first.”
@Brett: All things are *experienced* as scalar – even undeniably discrete things like “How odd/even is this number?” elicit gradient responses from people. (Armstrong, Gleitman & Gleitman, 1983; see https://omer.lingsite.org/blogpost-on-so-called-degrees-of-grammaticality/ for some discussion.)
That said, I suspect you’re getting caught up in precisely the confound I parenthetically addressed right under (4). Compare:
(1’) Which dish did you cook [before he devoured ___]?
(2′) This is the dish that I cooked [before he devoured ___].
(3′) This dish, I cooked [before he devoured ___]
(4′) * I cooked this dish [before he devoured ___].
Do you still find (4′) “marginal” in comparison to (1′)-(3′)? Or is (4’) decidedly worse than (4), in your judgment? If it’s the latter, then you simply failed to control for the confound I pointed out. To repeat: _read_ is only optionally transitive, and I was going for the reading one gets when one uses it transitively in (4). That reading, I think, is unavailable. (Grammaticality is a property of particular form-meaning pairings, not of strings.) In any event, (1′)-(4′) control for this by switching to an obligatorily transitive verb (_devour_).
As for (5), it is not missing a preposition. The point of that example is that it is illicit as a simple question about the object of the verb _reading_. The absent preposition ensures that the parse where it’s a question about a dependent of the noun _opinion_ is unavailable, forcing the reader/judger to try to construe it as a question about the object of _reading_ (which is not possible, and that impossibility is the point of the example).
@Omer Preminger: We evidently do not speak sufficiently similar English lects for this discussion to be useful. In particular, I find
your (1′) through (4′) all clearly ungrammatical.
Opinions of other [native] speakers?
Native (BritE) speaker here, happy to help. Unfortunately I’m not sure what I’m being asked to provide evidence on. Does the ‘___’ in the examples represent silence (an ‘understood’ NP, to use old-fashioned terminology)? Or is there to be a spoken “it”? (In which case, why isn’t that showing?)
If silence, I’m with Brett in finding all of (1′) through (4′) ungrammatical, and (4′) no less grammatical than the others — certainly the asterisk on that alone isn’t justified. (I wouldn’t say “clearly ungrammatical”. More borderline ungrammatical: I can imagine circumstances where I could give each of them some meaning.)
The earlier (1) through (4) equally just on the grammatical side of the border — including (4) with the sense of specifically the paper getting filed as not getting read. Another datapoint: If I switch position to “without first reading” I find (1) through (4) much more acceptable, although (4) seems rather telegraphic.
(5) is just plain ungrammatical without ‘on’ or ‘about’ following “opinion”.
Welcome, Omer, and I echo others’ thanks for your civil presence.
I do have to say that exactly the exercise you’re engaging in of posing sentences with no semantic context, and asking whether they’re grammatical (in itself a meaningless question to most native speakers), is exactly the sort of arid creek I criticise some branches of linguistics as having gone up ___.
consider sentences involving: (i) constituent questions; (ii) relative clauses; and (iii) topicalization
I don’t think that these groups of constructions are unconnected semantically though; might the common thread not be focus? (Admittedly, purporting to explain things as consequences of focus could be taken as obscurum per obscurius …)
I’m interested that you mention island phenomena, though. As Lameen remarked not long ago, that seems to be a genuine (and not very intuitive a priori) discovery of the generative school.
I certainly agree that both what falls inside my (parochial) definition of “syntax” and what falls outside of it are *definitely* worthy of study
Fair enough.
Here’s a much more lively example, not concocted:
Although this is from a foreign-language publication, it’s English service is usually excellent. (AmEng, occasionally a little stiff.)
I find that last sentence unacceptable, as failing to express what I guess it means. I guess: “… cannot be allowed to keep happening.” (but that’s lost the sense of “cannot afford” — where it’s the party that’s bearing the cost.)
The example exhibits deleted elements and ‘parasitic gaps’. Analysis please!
why all three “constructions” in (1)-(3) would behave alike: they behave alike because there’s only one mechanism they could be based on,
Just no: they could be based on three different but similar mechanisms. Why are they similar? (is then a reasonable line of enquiry) Perhaps for historical reasons: three incongruous mechanisms have evolved to the point they seem similar.
More generally, why does English (especially) have such a large vocabulary of words near-identical in meaning — typically one from Germanic roots; one from Norman/Romance roots? Because being ‘efficient’ is not a desideratum. Why does it have ‘native’ constructs from Germanic, and constructs carrying similar meaning calqued from Latin?
Hypothesising that languages are in some sense ‘efficient’ in use of syntactical resources or tend to “a paucity of means” is exactly what’s been wrong with Chomsky from the beginning. Languages are a glorious mess, they revel in being idiosyncratic, like a dog rolling in new-mown grass.
It’s only the grammarians who want an ‘economy of description’. That’s fine, but I reject the “only one … could” of “only one mechanism they could be”. The alleged necessity there isn’t saying anything about a language: you’re talking only about your methodology.
IIRC the alleged necessity of passive clauses being generated from active — the motivation in Syntactic Structures — has now been called into question: there are passive clauses for which there’s no corresponding active or no possible ‘by …’ agentive. So the ‘economy of description’ for clause constituents falls apart.
This is perhaps where I should mention that the kind of data I gave in (1’)-(4’) have been subjected to large-scale quantitative study with native speakers of (American) English, in the work of Jon Sprouse and his colleagues. (Though, admittedly, not (1’)-(4’) themselves; I didn’t have time to go look up the originals.)
Here’s another batch, intended to demonstrate the same thing. Curious what you all think:
(1’’) Which products did they examine [without ultimately purchasing ___]?
(2’’) These are the products that I examined [without ultimately purchasing ___].
(3’’) These products, I examined [without ultimately purchasing ___].
(4’’) (*) I sampled these products [without ultimately purchasing ___].
@AntC: You seem to be arguing against a claim I never made. I never mentioned “economy” or language being or not being “messy.” I said that *this specific phenomenon* – if it turns out to be real – can be explained if there is only one underlying mechanism for associating items at the beginning of the clause with a position inside it (e.g. the object position of _examine_). That is not an appeal to “economy”; it is merely an attempt at an account. You would be surprised, it seems to me, to find how much I agree with you re:actual appeals to “economy” – see, e.g., footnote 1 here: https://omer.lingsite.org/blogpost-how-minimalism-hijacked-modularity/
Regarding context, I fully agree. I assumed some shortcuts were okay in the comment section of a blog, though. My comments are plenty long as it is 🙂
Your “non-concocted” quote does not contain a parasitic gap, as far as I was able to find. I too find the last sentence of the quote unacceptable, but a parasitic gap is a very specific thing (a missing element inside an adjunct clause co-construed with an element outside that adjunct clause) and the sentence in question contains no such thing. We can disagree, but let’s at least set the terms of what we’re disagreeing *on*.
Lastly, one of the things that I think we do disagree on relates to the term “concocted.” As I discuss here – https://omer.lingsite.org/blogpost-ecological-validity/ – I think “concocted” examples are a better way to investigate language than “non-concocted” ones. I take it you disagree, which is your prerogative of course. I’m only bringing this up because the reason I hold this position, which you can read about in that (short!) blog post, might surprise you. I hold this position because… I believe language is a glorious mess.
@Omer P.: In the latest batch, is the switch of verb in 4” deliberate? “I examined these products [without ultimately purchasing __]” seems neither more nor less acceptable to my ear than the version with “sampled.”
FWIW I find 1” and 2” much more natural sounding than the parallel examples in your earlier two batches;* 3” is weird-sounding in isolation, but I could see it sounding natural in a particular discourse context that was set up in the preceding sentences and would expect it, if uttered aloud, to have some specific intonational cues to mitigate the weirdness.
*I agree with Brett on the second set; on the first set I suspect the problem may be that the example-concoctor has some idiomatic or occupation-specific sense in mind of what it means to “file” a “paper” (in which it would be somewhat unexpected and thus worthy of note for the filer not to have read the thing filed) that I am not grokking because it’s not a usage with which I’m familiar.
@J.W. Brewer: Oh, the switch from “examined” to “sampled” was an oversight. Feel free to switch it back to the former, so that the example matches the others in the set:
(4’’) (*) I examined these products [without ultimately purchasing ___].
Thanks for catching that.
OK, to be clear I agree with the * on 4″. But what about (assuming “these products” have already come up in the discourse to provide an antecedent for the pronoun):
(5″) (?) I examined them [without ultimately purchasing ___].
That one for some reason gets even more acceptable for me if I swap in “buying” for “purchasing.” Maybe the construction feels less acceptable in a more formal register, and “purchase” is a more formal-register word than “buy”?
I believe language is a glorious mess
Preach it, brother!
As I discuss here – https://omer.lingsite.org/blogpost-ecological-validity/ – I think “concocted” examples are a better way to investigate language than “non-concocted” ones.
Thanks for that; I continue to disagree, but for the first time I feel like I understand the argument in favor of such examples (which so irritated me during my enforced study of generative grammar).
The first thread that I ever read here included commenters from the generativist side, and it was a good thing.
There used to be a commenter named Norvin who made useful contributions; I think this (2014) was his first appearance.
Two (okay three) thoughts on concoction:
1. If you do the sort of work where contrasts between grammatical and similar-but-ungrammatical sentences are key to illustrating your hypothesized grammatical rules, you can’t necessarily rely on an adequate supply of the latter turning up in the wild. People do make mistakes and say ungrammatical things all the time, but you can’t necessarily expect them to make a specific mistake that conveniently violates the rule/pattern you’re trying to identify/explain.
2. That doesn’t mean you can’t start with a found-in-the-wild (or at least highly plausible/natural-sounding) sentence as your grammatical example, and then just modify it enough to get your contrary ungrammatical example. E.g., McCawley contrasted the grammatical “His arrogance was exceeded only by his ignorance” with *”His arrogance was exceeded only by his bicycle.” The latter is weird, but it might take considerable time to come up with a non-weird-yet-definitely-ungrammatical example to illustrate that contrast.
3. That said, the tendency of some writers to use concocted examples that are “wacky” or “zany” (McCawley was a serial offender here, e.g. the perfectly-grammatical “Waldo keeps molesting sleeping gorillas”) is something I disliked as an undergrad and don’t like any better all these decades later. I think my objection is (a) that it feels like a rather desperate attempt by a grown-up authority figure to hold the attention of stereotypically surly and tuned-out students by playing the clown, which is not IMHO a good pedagogical style (and if McCawley was in fact trying to amuse his fellow supposedly-grown-up scholars rather than immature undergraduates, so much the worse); and perhaps more substantively (b) some examples so wacky/unnatural as to be what you might call pragmatically ill-formed (the “My hovercraft is full of eels” type of sentences) seem likely to be psychologically distracting in a way that may undermine the validity of whatever judgments you get about their grammaticality. But perhaps there are “jabberwocky” examples, to use Dr. Preminger’s phrasing, that don’t create that risk, though, because they are artificial without being zany?
The “cook/devour” sentences are so obviously unacceptable that I wondered if Omer had accidentally copy/pasted the wrong thing. Compare the three versions, the difference should jump out:
Which paper did you file [without reading ___ first]?
Which dish did you cook [before he devoured ___]?
Which products did they examine [without ultimately purchasing ___]?
One of these things is not like the others. One of these sentences has a different subject in the subordinate clause from the main clause (you … he), *and* a different verb form, *and* it lacks the time adverb that the others have. The other two are OK, the “cook” sentence is absolutely not. To make it more parallel, try something like “Which dish did you cook before later devouring?”
Omer, you’re not on deadline here, go ahead and look things up. Are you really claiming that your “cook/devour” sentences are acceptable? At least four Americans (Brett, JW, me, my spouse) and one British-born (AntC) are here to tell you they’re not even close, so please do look up the reference and tell us where and when exactly this butterfly was collected, and whether it has the pattern you’re saying it has, because to us that’s an extraordinary claim.
(A bit hypocritically, I’m typing this in an airport. See you late tonight or tomorrow.)
Yes, I think it was Norvin.
—
That a researcher came up with a line (say: “we we we what we”) is already a fact and can be studied. Perhaps there are people who study exactly this: human creativity in production of nonsensical lines. I do not know linguists who study works of other linguists from this perspective.
Meanwhile another fact, that an utterance was actually uttered in the wild is usually well within the scope of professional interest of many linguists, irrespectively of listeners’ reaction.
Artifical examples are used for studying a different thing: people’s reaction (grammaticality judgements in particular) rather than usage.
I suspect some people simply do not believe either in grammaticality or in that it can be studied.
Such people do refer to the notion, but there is a consensus between me and any other Russian I know that nominative direct object is “wrong”. I’m confident that I do not use it, that I do not hear it. If I hear it, I will be surprised. I know that it is common in many Russian dialects, and I am also confident that if Leo Tolstoy and Pushkin were using it I would use it too.
So we can refer to this consensus by “wrong” or by “ungrammatical in our idiolects” or by an asterisk, and remain fully agnostic with respect to existance and nature of “grammaticality”.
Such people do refer to the notion, but there is a consensus between me and any other Russian I know that nominative direct object is “wrong”. I’m confident that I do not use it, that I do not hear it. If I hear it, I will be surprised. I know that it is common in many Russian dialects, and I am also confident that if Leo Tolstoy and Pushkin were using it I would use it too.
Can you give me a (dialectal) example? I’m not sure what you’re talking about.
LH, google “корова доить” (in quotemarks!)
The search engine immediately offers me a collection of studies of “номинативный объект” – but I am not ready to recommend any particular work…. I also do not remember its exact fuction (compared to ACC).
I hope it won’t offer something different to you:/ The first of them is about northern dialects but it is not just north and there are examples from Old Russian (and I think Baltic langauges?) as well.
I agree that ktschwartz’s alternative “Which dish did you cook before later devouring” isn’t nearly so weird as what it is offered to replace. It’s maybe a bit pragmatically odd in that the usual purpose of cooking something is so that it may be later eaten. (If the point were supposed to that the cook was expected to serve the dish to, or at least share the dish with, others but instead devoured it all him/herself, that’s not clear without more context.) So it still isn’t parallel to the others where there seems to be a background implicature that filing-without-reading and examining-without-ultimately-purchasing are noteworthy because at least modestly contrary to some sort of background expectation about how things usually go. When I was young and worked in the university library it would have been weird to ask me “What book did you reshelve [without reading ___ first]?” because reshelving-without-reading was the rule rather than the exception.
@ktschwarz: Sorry if I didn’t make this sufficiently clear in my earlier remarks, but the relative acceptability of (1) vs. (1′) vs. (1”) is not at issue here. What’s at issue is the relative acceptability of the (4(‘(‘))) sentence in each set when compared to *its own set-mates* (so: (4) compared to (1)-(3), (4′) compared to (1′)-(3′), and (4”) compared to (1”)-(3”)). Thus, for example, you may very well be right that switching (the reference of) the subject between the main clause and the adjunct clause degrades the overall acceptability of the sentence. But this factor is held consistent across the sentences within each set, so it shouldn’t be at issue.
Now, my claims about *those* comparisons (i.e., (4) compared to (1)-(3), (4′) compared to (1′)-(3′), and (4”) compared to (1”)-(3”)) might still be empirically wrong, of course. But I want to be clear that those were the only claims I was making. And I’m definitely taking seriously the judgments provided by the Americans and the Brit – that’s the main reason I attempted to construct the second and third data sets, in the first place!
This goes to a matter that often gets confused: grammaticality, a theoretical construct, is binary; acceptability, a behavioral measure, is gradient. By hypothesis, grammaticality is but one of the factors that determine acceptability. Of course, the only way to conduct serious scientific inquiry is to (attempt to) fix all the other factors that feed into the mix and see if manipulating predicted grammaticality yields the predicted effects on acceptability. But note that no such attempt was made (nor claimed to be made) with regards to comparing the sentences in (1-4) to those in (1′-4′) or to those in (1”-4”). Certainly, a myriad of factors change when we look *across* the different sentence sets, and you’ve correctly identified (what I think) is one of the important ones, namely the identity or non-identity of the embedded subject and the main-clause subject. And so I see no reason, a priori, to expect acceptability to remain constant across, e.g., (1), (1’), and (1”). What *is* relevant is whether, e.g., (1”-3”) are more acceptable than (4”) in your estimation. (I could still be wrong about that, of course; but that is at least something I did claim.)
Part of the responsibility for this confusion rests with generativists, of course, because we use the notation ‘*’ (and its absence) in inconsistent ways. Sometimes, it is meant to indicate predictions about *grammaticality* (which, as noted above, can only be translated into predictions for *relative acceptability* under controlled conditions). Other times, we used it as a mark of *relative (un)acceptability* for sentences within a demarcated set of stimuli (which is what I was doing, above). Clearly, it would have been better had we come up with two different sets of notations for this. But I assure you, no one in their right mind thinks that grammaticality translates, all by itself, into absolute acceptability (which would be equivalent to the claim that any sentence ever presented with a ‘*’ would be rated lower than literally every single sentence ever presented without one). This is why, e.g., Sprouse and colleagues test *contrasts* (of the kind I was trying to construct, above), not sentences in isolation. See here for some recent reviews:
https://www.jonsprouse.com/papers/Sprouse%20and%20Almeida%202018.pdf
https://www.jonsprouse.com/papers/Sprouse%20et%20al.%202013.pdf
Lastly, to answer your request about experimental support for the theory of parasitic gaps, this paper does a (quantitative) acceptability rating experiment to pre-test the stimuli for the more complicated question that lies at the center of the paper; but the pre-test basically confirms what the generative theory of parasitic gaps predicts:
https://www.jstor.org/stable/4490269
LH, google “корова доить” (in quotemarks!)
Thanks! That’s weird, I’d never run across it. I found the paper “Надо корова доить! Номинативный объект в севернорусских диалектах” by Roman Ronko.
He refers to a “Потебня 1874” that isn’t in the bibliography at the end 🙁
Олекса́ндр Опана́сович Потебня́
# In 1874 he defended his doctoral dissertation entitled Notes on Russian Grammar (Russian: Заметки о русской грамматике). #
Yes, it is the year of the original edition (which is online).
Look at the cover picture of this 2015 reprint. That’s a wonderful image of the effect Russian vowels have on me – mesmerizing. Until now I have referred to them somewhat cackhandedly as “operatic” in an appreciative sense (although I hate opera).
Cologne is currently busting at the seams with what I presume are Ukrainian refugees. A few days ago I talked briefly with another dog owner I hadn’t seen before. His German vowels came out of that conch shell.
Oops, that’s not a reprint, but a book with the same title.
Потебня А. А. Из записок по русской грамматике. Т. 1—2. — 1958
Oh, Ronko seems to have defended «Номинативный объект в древнерусском языке и севернорусских диалектах в ареальной и типологической перспективе» in 2018. (https://iling-ran.ru/web/ru/theses/ronko)
I am not sure I will like his analysis, but hopefully it contains an overview.
—
What bothers me is “В связи с тем, что на данном этапе у нас нет достаточного количества данных о номинативе объекта в южнорусских диалектах, мы ограничимся замечанием о том, что он там действительно присутствует и приведем несколько примеров: …” in the article referenced above by LH.
He does not have enough data about its use in southern dialects. I wonder why:/ Moscow university sends students to Arkhangelsk oblast (cf.) but it is not the only university in Russia…
@Omer P.: I think the issue may be that in your cook/devour set of examples, the 1-3 options are So Wrong that I (speaking only for myself) can’t really form a coherent judgment as to whether option 4 is Even Wronger. Perhaps that is some evidence that 1-3 are not merely low-acceptability on a gradient, but ungrammatical in a binary way? At least in my variety of English, which is enhanced by the fact that I think I may have an unusually small “grammatical-but-unacceptable” zone. I can identify a proposed sentence as “grammatical but not something I’d personally ever be likely to say” for any number of sociolinguistic reasons w/o thinking it unacceptable in the sense that “I’d be shocked to hear any L1 Anglophone actually say that, even though I might concede it is not syntactically ill-formed in a technical sense.” Although with weird enough example sentences I suppose the latter reaction could be elicited more frequently if that were the researcher’s goal?
I think the issue may be that in your cook/devour set of examples, the 1-3 options are So Wrong that I (speaking only for myself) can’t really form a coherent judgment as to whether option 4 is Even Wronger.
Now that I’ve actually looked at those specific examples again, I entirely agree with this description, and I’m actually unconvinced that option 4′ (as opposed to 1′-3′) is not in fact Slightly Less Wrong. They all seem about the same level of very-bad-and-not-sure-if-technically-ungrammatical. (But of course I’m not a Native Speaker.)
I do agree that perhaps example sentences to the effect of “the gloky kuzdra budlated the bokr…” (replace with your preference of manufactured words) might actually provide better judgements in some cases, because the semantic reasons for unacceptability would no longer be relevant.
Vaguely related: “No, you cannot say that, because this is not something that a lion does!”
[In the original Russian that I’ve been told this story in: “Так сказать нельзя, потому что лев этого не делает!”]
By the way, having gone and read the thread linked to above, I can now confirm that the Norvin in question is in fact one of my professors from back in my grad school days (‘06-‘11)! I guess I’m not all that surprised… the world of generative syntax is, at the end of the day, pretty small.
@Omer P.: Well, the gen. syn. world is slightly larger than one might have imagined, because I was skimming your website and thought you mentioned that self-same Norvin, of whom I am aware solely because of his occasional presence on this blog, as a mentor (which seemed to make sense because of where you went to grad school), but on closer review it turned out I was confused because you were instead praising a different generativist named Norbert. (Neither is what you would call a super-common name in the U.S. these days, but Norbert is, or at least was in prior generations, the more common one – one of the 250 most common names for US-born boys born between 1915 and 1932 before sliding downwards until it dropped out of the top 1000 after year-of-birth 1971 – but I knew a guy with that name a year or two ahead of me in high school and I’ve never met a Norvin in the flesh.)
@Omer … – I think “concocted” examples are a better way to investigate language than “non-concocted” ones.
@Hat Thanks for that; I continue to disagree, but for the first time I feel like I understand the argument in favor of such examples
I’m not persuaded. What is “short” about the post is it’s short on sense.
Omer’s examples of science are all from the physical sciences: the LHC, Brownian motion.
You’d only think those are comparable if you’re operating under a belief that language is somehow a special/isolatable human capability independent from cognitive abilities. The sort of thing that could only be explained by a specific gene.
There’s no analogous ‘ecological’ setting in which you can place humans without them bringing their culture and memory. There’s no machine you can train/configure to give ‘pure’ linguistic acceptability judgments, as you can configure the LHC to operate in vacuum with monstrous magnetic fields simulating point sources of particles.
Let’s have some examples of anthropological/cultural studies abstracted from ‘ecological validity’.
The idea that ‘Acceptability judgments appear to be reliable, robust, and replicable’ (the subhed of the cited paper) are getting ripped to shreds on this very thread before the ink has even dried.
The primary conceptual problem highlighted by this, though, is probably the notion that there is such a thing as (e.g.) “English grammar” that can be described unambiguously and retrieved from any speaker of “English”. Quite false as long as we look closely enough. Variation exists at all levels of language between all speakers.
Where this creates trouble for generative (or also other more theoretical) ventures is whenever researchers overgeneralize a phenomenon of some speakers as one of all, sometimes from nothing else than introspection, and proceed to claim that the phenomenon is to be explained by it being theoretically unavoidable. Practically nothing is theoretically unavoidable. It may well be the case that currently in English e.g. Norbert’s examples do, indeed, have a common gap property in sentences of type 1–3 but not 4. Maybe even all known languages do. But will the god-of-generativism step down from the heavens to smite me if I purposefully craft a conlang where this gap property holds for 1 and 4 but not 2 and 3? Probably not. Will this conlang remain humanly learnable? Probably yes. With enough effort, this could be even tested if we are really curious about this.
Functionally and historically-oriented linguistics has well-established machinery for explaining (indeed predicting) phenomena that are rare but possible, i.e. appearing impossible from small samples. In this example case, too, I expect the real explanation to be something to the effect that some functional commonality ties together sentences of type 1–3, which then is likely to put them on a shared path of diachronic development, to the exclusion of type 4 — i.e. sentences of type 4 are perfectly possible but just evolutionarily disfavored.
(I would stress also that that’s “and historically”, not “or historically” — “synchronic functionalist” analysis is at its purest also a dead end that provides no real explanations of most phenomena in language, even if we can perhaps attempt some with furious enough handwaving.)
Generative linguistics’ main contribution to this issue, on the other hand, seems to me to be various general-purpose excuses for claiming that something that is known to be possible is actually still impossible, and the data showing it to be possible is just an illusion that can be disregarded. Not a good look. As long as this continues, the field is irredeemable pseudoscience in my eyes. This is not the only disagreement that I find with generative linguistics, but it’s one that I repeatedly run into early on in an argument, and which usually makes me throw away in disgust whatever paper or textbook I was trying to get into.
Adding to the tragedy is that this surely should not be even necessary to the de facto core of generative grammar; it just requires stopping treating overshooting theories to the effect of Single Unified Platonic Grammar Of English (or, indeed Of All Language) as the only proper research output. Butterflies are not the enemy.
Is this really the level of discourse we want to have here? Declaring opinions that one disagrees with to be “short on sense”? Announcing that a paper whose claim is *quantitative*, and whose evidence for that claim is *quantitative*, to have been “ripped to shreds” based on the opinions of five people on a comment thread? This is not about agreement or disagreement, mind you. I seem to have convinced the proprietor of this blog of exactly nothing so far, but that hasn’t precluded a contentful exchange between the two of us.
I’m an interloper here, so far be it from me to presume to dictate to veterans of this space how they should conduct themselves. The little I can control, however, is what kind of exchanges I do and don’t spend time on.
Is this really the level of discourse we want to have here?
While I understand your discomfort, this is, after all, a blog and not an academic conference. I think “short on sense” is pretty mild as online discourse goes, and if you’ll think back to those less hospitable online spaces you mentioned, I’m sure you’ll agree. You are, obviously, in a basically anti-generativist space, and you can’t expect to convert many people, but I am benefiting from seeing your views laid out in a user-friendly way, and I’m not sure I’m the only one. Try not to read skepticism as hostility, is my advice.
Another strategy open to anyone in principle (your rileage may vary), is to conclude that few people will ever see things your way, and find a paying job in some other area. That’s how I ended up in IT. Nobody wants to hear about Luhmann and other Great Lights of Advanced Thinking, so I just snipe at the linguists.
Traditionally, enlightened beings retire to the mountains or deserts, and say nothing for years. I suspect it’s because they’re tired of being misunderstood, and prefer to sulk. I ended up there too. There’s just no balm in Gilead, nor toilet paper on the supermarket shelves.
I like Pystynen’s comment best so far, for content and style of discourse.
I want to explore another direction. There is a great deal of phenomena subsumed under acceptability, which are qualitatively different; a scalar measure of acceptability does not suffice to account for them. The main divide is between production and perception. GS concerns itself with production, but acceptability judgments depend on mechanisms of parsing and semantic interpretation, which GS has never concerned itself with. Thus “I wants this” is unacceptable in the sense that a speaker of standard English would never produce it, and it would surely be marked with an asterisk. But it is easily parsable and interpretable. There is some fascinating work done on parsing by psycholinguists; I’m very partial to Fernanda Ferreira’s work on “Good Enough” parsing. But none of that figures in syntactic models of what an acceptable utterance is.
Now in phonology and phonetics the equivalents, namely the production and perception of sounds and the mismatch between them, are recognized as fundamental forces in shaping the way phonologies evolve (Blevins’s Evolutionary Phonology program is the purest distillation of the concept.) But GS does not have that second half and I think that is a fundamental flaw that hampers it right from the beginning.
GS concerns itself with production, but acceptability judgments depend on mechanisms of parsing and semantic interpretation, which GS has never concerned itself with.
It takes two to talk.
I’m not sure I understand where Y is going. Obviously there are ungrammatical sentences that are easy to parse/interpret and others that are more ambiguous and/or nonsensical. But are the former more “acceptable” in a useful sense of “acceptable”? The point underlying your example is that the -s suffix that distinguishes third-person-singular verbs in (non-vestigial-subjective) present tense from the other person/number possibilities in (non-vestigial-subjunctive) present tense is generally redundant in modern English. Which means that neither adding it where it’s forbidden nor omitting it where it’s obligatory will interfere much with interpretation. But redundant features like that are ubiquitous in morphosyntax, and, given that speech often occurs in low-fidelity environments with lots of background noise, there’s an argument that some degree of redundancy is a feature rather than a bug. That our internal mental parser can easily disregard certain sorts of syntactic errors but not others is a true and interesting fact, but it’s not clear to me how that fact should inform the syntactic description of a given language or a more abstract theory of syntax in general.
Put another way, in most languages the various tacit rules that must be complied with to successfully utter a grammatical (or acceptable) sentence impose stricter limitations than “say whatever words you want with or without whatever morphological markers and in whatever order, but only up to the point where you’re no longer likely to be understood by another native speaker’s internal mental parser.”
Is this really the level of discourse we want to have here?
Please don’t be put off by this, Omer. It is a real pleasure to hear from a generativist who is so refreshingly free of the implication that their approach is the the One True Path; speaking for myself, at least, I think a lot of the annoyance I feel at Chomskyans (who are not all the same thing, of course, as your own original post makes very plain) is to do with such maximalist claims, and indeed to what you might call the sociology of that school. I strongly suspect that this has impeded me from learning some things of real value.
I recall that I myself (by no stretch of the imagination a generativist) provoked some mild derision here by innocently referring to “gaps” as a grammatical concept (not a very controversial position in itself, I thought …)
It takes two to talk
I don’t find this to be so; but if this is indeed a cultural norm, it may explain some of the looks I have been getting lately …
I’m relieved to learn that Omer isn’t actually claiming that “Which dish did you cook before he devoured?” is acceptable. As JWB said, the cook/devour examples (1′-3′) are already so far out that it’s meaningless or impossible to judge whether (4′) is even worse; the question seems bizarre, I can’t answer it. JWB is also right about the pragmatic weirdness of that set — something felt off about them, but I couldn’t articulate it. An example that would really be parallel to the others would be e.g. “Which dish did you cook without ever enjoying?”
What’s been ripped apart here is not the Sprouse paper itself, it’s the claim that your (1′-3′) are measurably worse than your (4′). As you say, that example isn’t in the paper, so I’m wondering if you overgeneralized from what *is* in the paper; did they test anything as overall-bad as that set? I’ll try to find time to read it. (And AntC clearly did not dispute whatever quantitative conclusions Sprouse presented, he disputed the far more general claim that “Acceptability judgments appear to be reliable, robust, and replicable”.)
I’m skeptical about this sort of acceptability test anyway since there are so many examples where speakers will judge something as “wrong” in isolation that they have no problem accepting in context, and even producing themselves. Labov didn’t just ask people which pronunciations were right or wrong, he elicited responses in context.
@ktschwarz,
Setting the empirical record straight: Acceptability judgments appear to be reliable, robust, and replicable is the title of the first paper by Sprouse and Almeida linked by Omer.
(link 1, the same that Omer posted and link 2 in case anyone needs their references)..
It is a brief commentary on another article, (this one).
it may explain some of the looks I have been getting lately …
Yes, walking around the veggie shop muttering to yourself about the price of tomatoes/red peppers/how avocadoes are never properly ripe in this bloody country despite being grown here, appears not to be within cultural norms.
I suspect it’s also beyond cultural norms to rush into a crowded seminar room; say/write “I cooked this dish before he devoured”; and rush out again. Then expect listeners to make any sort of judgment. (It’s quaint research subjects are called ‘native speakers’, because speaking is not required of them.)
We’d all do the Gricean thing upon hearing such sentences, of generously trying to give a coherent interpretation. Judging them grammatical/acceptable is no part of the culturally normal exercise.
Since Chomskyans focus so much on ‘generative syntax’, really the only way to organise an isolated experimental environment is to get speakers to spontaneously produce utterances with ‘parasitic gaps’.
I’m relieved to learn that Omer isn’t actually claiming that “Which dish did you cook before he devoured?” is acceptable.
Objects raise and they devour.
“I let the guests in without greeting, would I also let them out without entertaining?”
@Omer you can read about in that (short!) blog post,
@AntC What is “short” about the post is it’s short on sense.
Is this really the level of discourse we want to have here?
Yes the pun was lame. Appropriate responses would include continuing the banter with a lamer pun; or a witty quote; or a surrealer flight of fancy.
You believe language is a glorious mess. But I’m not sure you grokked my bit about the ‘reveling’. Despite processing language being right at the outer limits of human cognition, speakers always show an ability to step outside the languageing/go meta and play with language for the sheer fun of it. Perhaps posing bizarre sentences is some sort of play, but why so poe-faced about it?
Bizarre sentences in a research context are like some sort of bathing-machine carried about on dry land, rip’d untimely from their usefulness or cultural relevance. I’d be just astonished if “Acceptability judgments appear to be reliable, robust, and replicable” under those circumstances. I’d suspect the researchers to have tin ears.
(Like Hat, I was first exposed to the white heat of Chomskyanism in early 1970’s Undergrad course — mine jointly between the Philosophy and Linguistics departments. Whereas the Philosophy of Language was all Wittgensteinians agonising about How to do things With Words, Sense and Sensibilia and language games, meaning as use, the private language argument; what I remember of the Linguistics content — apart from the white heat — was the hopelessly amateurish clowns Katz and Fodor, and Lakoff’s lot kissing their toothbrushes and dreaming of falling in love with their typewriters. And then Chomsky — I did read Syntactic Structures, with its bizarre examples — advocating taking the meaning out of language all together. I’d have to say things went downhill after SS; I still own a copy of Aspects: it is mostly unread.)
Can we call it Chomskism and save a lot of syllables? (TIL that Anglos have to be Trotskyists instead of Trotskists like in sensible languages, but that’s still a syllable saved over Trotskyanists).
If Chomsky gets LH in Google search results as often as I do… That is very often.
(though my requests often include names and forms from exotic languages)
Google just need to learn that Chomskism involves Chomsky, can’t be that hard.
“I let the guests in without greeting, would I also let them out without entertaining?”
Sorry, that gets an asterisk. You have to add the object pronoun.
@lh
There are sentences quite close to this one, that are acceptable to me, i.e.,
—
I let the guests in without saying hello and subsequently let them out without serving any drinks.
—
But I suppose here it is clearly an indirect object and not a direct object that is being elided?
Yeah, that’s entirely different.
I mean, you can say “I said hello,” but you can’t say “*I greeted” (unless you’re Scots and talking about weeping).
@Lars: The google n-gram viewer reports that “Chomskyanism” and “Chomskyism” have been pretty evenly matched in frequency of use since the late 1990’s, after previously taking turns as the more popular variant. But when it comes to describing an individual adherent or needing an adjectival form, “Chomskyan” is the overwhelming favorite, with “Chomskyist” and “Chomskyite” hard to distinguish from the x-axis. Of course “generativism/t” have even more syllables …
Over the last half-century or so, the n-gram viewer reports “Trotskyist” as more common that “Trotskyite,” although I adhere to the latter. I believe the -ite form may be more of an endonym and perhaps perceived as more pejorative, but those don’t strike me as good reasons not to use it.
Germanic languages seem to be unusually fickle with object pronouns – and German is even more so than English: ich weiß nicht is not a complete sentence, you need das weiß ich nicht or in a pinch ich weiß es nicht.
Oh, wait, maybe that cuts right across Romance: Spanish no sé vs. Italian no lo so.
Ha, that one (ich grüßte) actually works in German.
I assume Danish is taking after German Trotzkismus/Trotzkist here, but then they spell the guy Trotzki because Russian.
FWIW, in Danish it’s totally unremarkable to say Jeg lukkede gæsterne ind uden at hilse. But then greeting someone is participial in Danish (hilse på nogen). The other part about entertaining is dicier because underholde does take a direct object, but at underholde can be undestood as “performing entertainment” so it’s not right out to omit the pronoun.
Doesn’t German also use “Trotzki” for the guy himself?
Leo Trotzki.
@LH, are “greeting the guests” , “entertaining the guests” wrong? I thought I have seen such phrases…
Those phrases are fine; they have an explicit object after the verb.
“I started dinner so late that I was greeting the guests while still making butterbean dumplings.”
Almost Prutkov “I started dinner so late that it ended as a breakfast”
@LH, sorry, I misunderstood you:( I was thinking about this:
“I would be leery of any syntactic argument that distinguished your (4) from (1–3) as operating under different mechanisms, because I find, “I filed this paper without reading first,” to be grammatical. However, it is awkward, in a way which I interpret as indicating that it is almost ungrammatical”
Omer described it as:
I thought that maybe “without reading first” works similarly to “without looking”. If so, I wonder if there is a border between verbs that can work this way and verbs that can’t, if the former class is closed and what unites the verbs in it…
For this reason I was trying different verbs: “…[verb] the door without opening” etc. “Greeting” is restricted to a certain genre and is quite unlikely to join the aforementioned class. Which reminded me about the difference between “buying” and “purchasing” (JWB above).
Yes, it was meant to require an object.
It was meant as 4”’, in other words. But I was fooling around and I was not sure if we need another variant of (4).
I* find, “I let the guests in without greeting, would I also let them out without entertaining,” fine, but only because I don’t perceive a missing direct object there. Unlike greet, the object of entertain is optional. So much for perfect intersubjective agreement on grammaticality.
* Of course, my judgements of grammaticality have been questioned by other native speakers. Moreover, I have, on occasion, used what were to me grammatical but unlikely statements with the knowledge that they would likely not be understood. (One time, while attending a meeting of a large organization, of which I was one of the officers, somebody who I didn’t know tried to shoo me out of one room, claiming I wasn’t permitted there. He refused to acknowledge that I was not under his authority. “Were I, would you,” I told him at one point, after I had given up trying to explain the actual situation.)
And speaking about nonsense, as always:
“What proletarians have nothing to lose?”
“Besides their chains!”.
(a Soviet riddle. In English they have nothing to lose but their chains, but but is ambigous).
in the spirit of the joke about opera singers where an error of one (“кто там в зелёновом берете”) triggers an avalanche of errors from confused others (“сестра моя…” “так ты сестрат!?”)
“Unlike greet”
@Brett, is the first half fine (for you) as well? If yes, why?
P.S. I wrote “Unlike greet” at first, but it seems the trick does not work with two-word quotes:(
I wonder if there is a border between verbs that can work this way and verbs that can’t, if the former class is closed and what unites the verbs in it…
This kind of analysis is the pits. It treats English as if it were a language of heptapods. You’ll need to become a heptapod to master it.
There’s still time to change course: nondum etiam sensus deperditus omnes.
I* find, “I let the guests in without greeting, would I also let them out without entertaining,” fine, but only because I don’t perceive a missing direct object there. Unlike greet, the object of entertain is optional.
Like drasvi, I’m confused by this. Since greet requires an object, how can you find the first part OK?
I apologise for any whiplash from a not entirely academic tone, but regardless I’m not going to sugarcoat my point about philosophy of science. If a supposedly scientific discipline subscribes to a method that empirical evidence that disagrees with a theory can be disregarded from consideration for no other reason than that it disagrees with the theory (data might be “just a production effect” etc. rather than “actually grammatical”, and such a claim will not need to be justified by any independent evidence), it has become unmoored from reality. Results by practitioners of any such method are fundamentally under risk of being arbitrary nonsense. Possibly not an especially massive risk: disciplines turn slowly and might be able to be re-moored with relatively modest effort. Agree also with AntC that any problem of robustness etc. is itself just in the (small?) gray zone in-between unambiguously grammatical / ungrammatical, in those phenomena that call for “particle linguistics”. But then if this is treated as key for resolving theoretical issues (already since Chomsky’s early career?), it is also possible that the repercussions could reach far.
Anyway, a further analogy, if you will (without insisting on the Construction Grammar viewpoint that it’s actually a full identity). Ambiguous gray zones are exactly how e.g. the lexicon works: some words are clearly in “the lexicon of English”, some are clearly not, but then for a gray zone of words they “are” in the lexicon only depending on the time, place, speaker, social context. The same hold for idioms, or quotes. Same even in much smaller / more discrete inventories like phonology. Same in semantics too, morphology, writing systems. So why not syntax? Lexicographers left the that’s-not-a-word attitude behind long ago in favor of close-to-reality descriptivism and I hope to see that some day “syntactographers” or “constructographers” will follow on their path. On which topic, I would argue that even “I wants this” is surely some part of “English”, even if it is tagged as only appropriate for child / jocular / L2 learner registers.
And even after the beating of descriptivism, the notion of “real English words” does not entirely vanish: it splits apart instead into a cluster of correlated-but-separate concepts like “core vocabulary”, “neutral in register”, “literarily standardized”, “phonologically integrated”, “etymologically native”. I expect something like this to be the reality behind “grammatical English sentences”, too.
…though unambiguously requiring a Polish pronunciation. Or perhaps a Ukrainian one.
(The t makes the o short, meaning [ɔ], which doesn’t exist in Russian; instead, the original has a pretty long [o] because it’s stressed, but the sequence oz is very odd from a German point of view… indeed he gets pronounced with [ɔ] in German, as the spelling demands.)
omnis (nom. sg. m./f.)?
Yes, omnis (III.11).
I like that.
Me too.
@Stu, but I am not teaching LH to speak my idiolect…
For JWB “without buying” is better than “without purchasing”, and I think for everyone “without reading first” is better than “without opening” (but I am not sure: in Russian both would be fine).
I wonder why because I wonder why. If you know a way to wonder differently, please share.
If you know a way not to wonder at all, I can do that. But why?
omnis (nom. sg. m./f.)?
That phrase nondum etiam sensus deperditus omnes is quoted in two different ways in the internet: sometimes with omnis, sometimes with omnes.
I’m sticking with the omnes reading, because of the “accusative of respect with the perfect passive particlple” explanation here.
You guys seem to think there is a “nominative of respect”. How else to analyze sensus deperditus omnis ? And which single sense, of the five meatspace senses, would it be that the poet has not yet lost ?
Of course it’s just possible that the poet, or one of his later editors, lost his sense of syntactical fitness.
If you know a way to wonder differently, please share.
Use analytical tools as crutches, and discard them as soon as you can walk. Even slightly sooner.
With languages my goal is to walk and run, not to become a crutch virtuoso.
Just imagine how hungry you would be if you could eat only with chopsticks.
That phrase nondum etiam sensus deperditus omnes is quoted in two different ways in the internet
I was going to say something stern about your offering “the internet” as the ultimate judge of these things, but I see actual scholarly editions go different ways on this (here, for instance, we have “omnes” in the text and “omnis” as an alternate reading in the notes). My Latinity is far, far too exiguous to permit me to even have a notion about which is better and why.
@drasvi, languagehat: Honestly, I can’t say for sure why, “I let the guests in without greeting,” seems grammatical to me. Obviously, a much more natural locution would be, “I let the guests in without greeting them.” However, something about the shorter version licenses the interpretation of “the guests” as the direct object of “greeting.” It doesn’t seem quite right that it’s a gap, and, “let… in without greeting,” lacks the parallelism of a compound predicate;* nonetheless, it feels rather like being an amalgam of the two is what makes it grammatical. (How non-Chomskyian!)
* Is it actually possible to order the component verbs in a compound predicate so that their common direct (or indirect) object falls between the verbs? It seems like it could happen, although examples are not so readily forthcoming. The more modifiers on the second verb, the better it tends to sound, although it’s not that simple.
Same here.
I didn’t offer “the internet” as judge of anything, but rather pointed out that it is inconsistent.
What I’m sticking with for the time being is the “accusative of respect” explanation, because it gives me something to work with. One has to make decisions before one can revoke them later, after prayerful consideration.
?He memorized a whole chapter of Heideggerian gobbledygook and regurgitated on command.
Change that “?” to a “!”, because “regurgitated” doesn’t require explicit specification of the regurgitatum.
“I let the guests in without greeting,” seems grammatical to me.
Too me two. Because there is a “parasitic gap” or “zero” there: “without [a] greeting”, that turns the participle/gerund into a noun – if only you let it. If we are to live with gaps and zeros, let us be generous with them.
There is no escape from sense.
“I wants this” would be perfectly acceptable in the variety of English spoken by Gollum, although I guess he tended toward the “royal we” so “we wants this” would be more likely?
I didn’t offer “the internet” as judge of anything, but rather pointed out that it is inconsistent.
But who cares whether it’s inconsistent or not? It’s like saying “I went out in the street and asked a dozen madmen, and they had different ideas about it.” The internet is utterly irrelevant on questions of accurate quotations. (If you believe “the internet,” all quotes are by Mark Twain, Wilde, or Churchill.)
Of course. I’m still in Inconsistency 101, we haven’t yet covered accuracy. That was never a concern of mine in this matter. Productive plausibility is my butt buddy.
I expect that from now on you will be rapping commenters on the fingers whenever they cite the WiPe.
Now now, we all know by now you don’t believe in reality, let alone accuracy. Don’t try your trick answers around here.
I do not expect that what I wrote would be taken as an accurate description of my views. But others will take it as they please, as usual. Cur quis non prandeat, hoc est?
That about exhausts my stock of Latin tags. Time to go shopping.
ich weiß nicht
Grammatical for me like this, without object pronoun, when used as an expression of doubt or hesitation (see also ich weiß nicht so recht…).
Actually, one reason why I liked that example is a song from the Russian audioplay Alice in Wonderland (by Vysotsky).
впустить или не впустить
а если рискнуть,
а если впустить,
то выпустить ли обратно…
(I didn’t comment on that ruse at the time — bigger fish to fry.)
Asking for judgments on acceptability modulo rejecting some interpretation seems exactly to tangle up meaning/pragmatics with syntax — even if the interpretation to reject is “somewhat odd”.
Given this Glass Bead Game is already concocting implausible sentences, just concoct a sentence not susceptible of the undesired interpretation. Or if you can’t do that, that should tell you something without needing to trouble a bunch of people.
Lexicographers left the that’s-not-a-word attitude behind long ago in favor of close-to-reality descriptivism and I hope to see that some day “syntactographers” or “constructographers” will follow on their path.
Thanks @J P, well put.
@Stu:
You’re right (naturally.) It’s an accusative of respect. The accusative plural of omnĭs is omnīs (with a long ī) or omnēs (which is a later form.) It agrees with sensūs, also accusative plural. It’s not nominative singular omnĭs here.
The construction is supposed to be influenced by Greek. I thought you’d like to know that. (That Propertius was well cultured.)
Archibald A. Hill (1961) Grammaticality, Word, 17:1, 1-10
about (1) Colorless green ideas sleep furiously. (2) Furiously sleep ideas green colorless
You’re right (naturally.)
Well, it’s a Pyrex victory at best. Apart from yourself, apparently no one here knew the Latin spelling rule ” ‘i’ before ‘e’ diachronically”. We were all operating in the dork.
“the child seems sleeping” is a strange example. Is not it something that was fine in the 19th century at least?
Yes, it is acceptable but seems old-timey in some way I can’t pin down immediately. Could be mock old-timey, but I don’t think so.
It’s ungrammatical for me. I’d say “seems to be sleeping.”
From Maxims and Epigrams, by Orlando Wright (1876), p. 89:
From A Timid Brave: The Story of an Indian Uprising by William Justin Harsha (1888), p. 79: “Yes, there is the little babe, and she seems sleeping peacefully.”
From Transactions of the Minnesota State Medical Society (1877), p. 72 (report on patient): “12 P. M. – Seems sleeping.”
A search on “seems sleeping” restricted to the 19th century turns up many such hits.
” “I let the guests in without greeting,” ”
Maybe there should be a borderline grammatical category for “Mistakes people sometimes make which they don’t always bother to correct since they’re almost OK and which in the course of time might become OK if enough people do that”.
Sounds like calling someone a spouse because they might get married one day.
The word you’re looking for is “fiancee”.
Or perhaps just “fling”.
.
No, I’m making a point about your “borderline grammatical category.” Calling something grammatical because in the course of time it might become OK doesn’t make sense to me.
I didn’t call it grammatical, I said that maybe grammar should have a borderline category. That sentence doesn’t look grammatical to me but I can easily see it becoming grammatical, or becoming a local variant. At which point anti-prescriptivism would kick in.
Predicting possible changes isn’t possible, but that seems to be a place where a change wouldn’t be that hard.I never would have thought of it if it hadn’t shown up here as a horrible example though.
Fair point.
Language change at least implies that borderline cathegories are common (unless we expect it to be quick and simultaneous across the community).
Also true.
Language change at least implies that borderline cathegories are common (unless we expect it to be quick and simultaneous across the community).
Depends on what you mean by “language change”. If a country is conquered by another, in which a different language is spoken, there is immediately a language border within that country: between the conquerors and the inhabitants.
A different kind of language change, often discussed here, doesn’t have such a border effect: those Great Vowel/Palatalized Velar Shift things. These apparently occur over a hundred or so years, sometimes spreading geographically. At least I’ve not heard any argument that they were quick and simultaneous – an explanation for such a phenomenon would require positing alien intervention, or a throat virus epidemic that deformed everybody’s velars.
I have no idea how such Great Shifts actually played out. I imagine the changes were picked up (imitated) gradually by some individuals from other individuals, but not all from all. The “borders” at each instant would be between individuals, not between geographic areas. The shift would be like a wave of foam with continually changing shape, each foam bubble enclosing an individual and forming a “border” to others.
“Bubblebath categories” might be less misleading than “borderline categories”.
I have have heard the Northern Cities Vowel Shift.I grew up in the Midwest preshift but have lived on the West Coast since 1964, more or less, But a friend of mine who also grew up there has stayed, and he has shifted. I even have heard the “Cities” part, because the rural people I grew up with did not shift, but my friend is urban.
https://slate.com/human-interest/2012/08/northern-cities-vowel-shift-how-americans-in-the-great-lakes-region-are-revolutionizing-english.html
@Stu, LH, DE, thank you. “The child seems sleeping” is Chomsky’s example of an ungrammatical construction. Google n-grams. Most books are poetry, but there is a semantical subtlety:
1848 : “the village below me seems sleeping”
1830 : “and whilst all the town seems sleeping in cold frost and fog”
The Motor, 1938: “Nestling below the hills Aberfoyle seems sleeping in a climate always warm , and indeed mildness is a feature of the weather here…”
I can imagine that “seems sleeping” better conveys impressions.
And there is:
The Oxford and Cambridge Grammar and Analysis of the English Language,
by George Gill, 1874, p 140:
“4. Many Intransitive Verbs take the same case after as before them; as,
He became a minister.
William seems a good soldier.”
revised and enlarged by the Rev. Charles Brooke, M.A., 1925:
This is not quite the same as John Emerson’s point, but my mom grew up in the Buffalo area before permanently moving away from Upstate New York after graduating from college in 1960, and I doubt (although I have not done a lot of A-v.-B fieldwork) that her vowels have Northern-City-shifted to anything like the same degree as those of many of her high school classmates who stayed in the area.
there is a semantical subtlety
Actually, I do myself find “seems sleeping in cold frost and fog” OK, even though I wouldn’t say “the child seems sleeping.”
I think it turns on the fact that in “the child is sleeping”, “sleeping” can be taken either as forming part of the present continuous or as equivalent to “asleep”, effectively as a predicative adjective. Adding an adjunct that can be taken with “sleeping” in an adjectival way makes the second interpretation easier. On the other hand, adding a manner adverb makes it impossible: *”The child seems sleeping badly.” (Just as you can say “the town seems asleep in the cold frost and fog” but not *”the child seems asleep badly.”)
How do speakers who accept “The child seems sleeping” feel about parallel constructions where the verb+ing form doesn’t lend itself to being taken adjectivally so readily, like “the child seems crying”?
If you don’t find that one acceptable, maybe Chomsky (not for the first time) managed to pick a particularly bad example for his diagnostic sentence. (The man just has no feeling for language …)
The child seems peacefully asleep.
The child seems peacefully sleeping.
The gentleman who wishes to be thought serious rather than a fop adopts the motto “Esse quam videri.” The comparative-philological scholar, by contrast, investigates the varied syntactic forms of permissible arguments for videri …
How do speakers who accept “The child seems sleeping” feel about parallel constructions where the verb+ing form doesn’t lend itself to being taken adjectivally so readily, like “the child seems crying”?
From Legends of the War of Independence by Thomas Marshall Smith (1855), p. 156: “He must have done, or be about to do something very bad, as he seems crying most piteously already.”
From Good Taste, Bad Taste, and Christian Taste by Frank Burch Brown (2000): “The baby seems crying not so much for itself as for her. ”
With “going,” Byron wrote “My hair though not gone seems going,” and in Daily Life in 18th-century England by Kirstin Olsen (1999) we find this 1729 quote:
With “coming”: “Slowly, slowly, through the countless ages, some one’s grand dream seems coming true”; “A flood of that miserable heresy, universalism, seems coming in upon us”; “a man with the soul of that beautiful white Christ which seems coming out of Russia”; “Firing now seems coming from direction of old harbour”; “neither can the Cambridge Town and County Cricket Club prevent the fate that evidently seems coming upon them”; etc. etc. (Many more; put “seems coming” in Google Books search and see for yourself.)
The child seems peacefully sleeping.
OK; so it’s not that it’s a manner adverb …
Take two …
Perhaps it’s continuous versus simple present:
“The child seems to be sleeping badly” = “It seems that the child is sleeping badly.”
“The child seems sleeping badly.” = “It seems that the child sleeps badly.”
“The child is asleep” is equivalent to “the child is sleeping”, but not to “the child sleeps”, except in old-timey and poetic registers in which the simple present does duty for the continuous present, and in things like the “stage directions present” of e.g. “He shoots! he scores!” or “Godot enters stage left.”
So my substituition test with “is asleep” for “is sleeping” might be picking up that, rather than the syntactic status of the word “sleeping” itself.
This would actually chime with drasvi’s suggestion, that such uses of “is sleeping” convey “impressions”: they are descriptive rather than dynamic. The addition of picturesque background details about mists would go with descriptiveness rather than dynamism, too.
“Godot enters stage left.”
From the little-known sequel “Godot Finally Shows Up.”
A moment later Godot exits again, pursued by a bear (and you thought it was a boring play.)
He seems asleep.
A sleeping dog .
A dog, asleep. .
An asleep dog (??)
He seems sleeping .
I let them in without greeting.(??)
I let them in without a greeting.
I gave them a drink without asking.
These are pretty abstruse grammatical rules.
Again my point: production grammar is evidenced through corpus studies, from which we can see that “the child seems sleeping” is producable, whereas Chomsky’s perception grammar does not consider it acceptable. In other words, acceptability judgments are not a reliable source for understanding production grammar.
To take another example, when in Down by Law Tom Waits asks Roberto Benigni, “For why are you in this prison put?” the sentence purposefully does violence to English production grammar in several ways; the movie’s audience’s perception grammar recognizes it as such, but can easily parse it and interpret it unambiguously.
Strunk and White, notoriously, fail to practice what they preach (as Pullum puts it). They produce sentences which to them seem ungrammatical to produce.
Sounds like calling someone a spouse because they might get married one day
When someone was explaining to me that the polite way in Kusaal to address a grown woman whose name I didn’t know was m diemma “my parent-in-law”, I expressed some surprise that this was the convention even if the woman was actually younger than me and unmarried. He explained, “Ah, yes, but who is to say that when she is older she might not have a daughter you could get married to when she grew up?”
To me, the contrast between “I let them in without greeting” and “I gave them a drink without asking” sounds like a pretty good demonstration of why syntax and semantics can’t be compartmentalized as much as the GS’ers would like them to be. I hope Omer Preminger comes back to give his take on this.
Now John Emmerson got into poetry. Lemme make a few “improvements”
He seems asleep.
A sleeping dog .
A dog, asleep. .
An asleep dog (??)
He seems a-sleeping.
I let them in without greeting.
I let them in without a greeting.
I gave them a drink without asking.
These pretty abstruse grammatical rules
Are tasking.
@Y: Unfortunately, I think he was scared off by our gruff, ruthlessly empirical attitudes.
“It takes two to talk” …
A friend fo mine once posted a text in Russian online for corrections. There are specialized sites where native speakers correct essays and she just began learning Russian. Usually essays in English usually do not receive any corrections because of the obvious issues with supply and demand. Essays in Russian in turn get correct by some 4-5 people who rephrase everything and mark everything with red ink. It was driving her mad, and as these corrections where different, she asked me why. She calqued a construction from her native language which does not have a parallel in Russian, but there are three similar Russian phrases and three (out of four) correctors used these three phrases. Two are common, one a bit dated/literary. I explained it to her and she asked which correction is better. I said: if we do not count this phrase, it is the the one that received several downvotes from others (the lady who used the dated construction). “But why did they downvote it?”. I said, perhaps they do not read books as often and are not familiar with the construction (which for correctors means “wrong!”). It did not comfort my friend at all.
Strunk and White, notoriously, fail to practice what they preach (as Pullum puts it).
As I point out in the Reviser’s Introduction to Strunk & Cowan, there is no reason why they should: the “little book” is not a freshman composition.
They produce sentences which to them seem ungrammatical to produce.
Not ungrammatical, for the most part; their objections are about grammar in the technical sense. Strunk doesn’t think student body is an ungrammatical compound, he just doesn’t like it and wants people to write studentry instead.
Just noticed that Geoffrey Pullum discusses this very same Chomskyite effusion “the child seems sleeping” here
http://www.lel.ed.ac.uk/~gpullum/realism.pdf
with his customary lucidity. (Pullum’s not Chomsky’s. Don’t be silly.)
That’s an invigorating read, as is to be expected. I wonder whether it’s a typo or a thinko in “completely car-before-horse backward”?
Aha, so Chomsky here calls it “semi-grammatical”.
Compare:
Possibly neither. The horse came before the car. Difficilior for the win !
“Hoarse before catarrh” is another reversal of the usual order.
The horse came before the car
It’s an inheritance from Proto-Indoeuropean.
@DE, do you actually read “sleeping badly” as “sleeps badly” (which I take to mean: the child is a poor sleeper)?
I certainly don’t.
Are there studies of acceptability among less WEIRD (or else LOL) people?
(to quote Lameen: “Most Algerians will tell you that Algerian Arabic has no rules, but that won’t stop them from looking at you funny if you try saying something like: …”)
LOL (Literate, Official, and with Lots of users) is a linguistic substitute for WEIRD (Western, Educated, Industrialized, Rich, and Democratic) proposed by Östen Dahl (How WEIRD are WALS languages?).
this very same Chomskyite effusion
Somebody above complained ‘Chomskyanism’ is too long, can we please reduce to ‘Chomskyite’.
I used the -anism form modelled on Marxianism vs Marxism — so as to separate all the hangers-on corrupting the original thought.
What Pullum’s analysing I think justifies this separation. Katz is a Chomskyite, not a Chomskyanist, also a bumbling clown. (I’m not defending Chomsky, of course, and C might have espoused some of these opinions at times; but he moved away from them pretty quickly — admittedly retreating into clouds of the ineffable.)
How WEIRD are WALS languages?
https://www.eva.mpg.de/fileadmin/content_files/linguistics/conferences/2015-diversity-linguistics/Dahl_slides.pdf
Obviously I agree with the general drift of this presentation, as an enthusiastic publicist for a particular distinctly non-WEIRD/LOL language; still, as so often, it’s easier to recognise the problem than to come up with practicable answers (apart from the perennial “More research needed.”)
I like “Old-World Supersaharan” as a term …
Chomschism.
Are there studies of acceptability among less WEIRD (or else LOL) people?
Every grammar of a non-WEIRD/LOL language depends on such study …
Kusaal speakers are certainly happy to make acceptability judgments. Dispreferred usages tend to be attributed to Mooré influence, in some cases probably actually correctly, though the thing seems to have become a sort of established trope for “foreigner speak.”
My informants were also happy to criticise the grammar of speakers younger than themselves. I strongly suspect that this is a human cultural universal …
I was very lucky in having several informants who were very good at imagining contexts in which an utterance which seemed at first blush to be ungrammatical would turn out to be acceptable after all: the same sort of thing as Hatters have been doing here with English, in fact.
Like the consultant for a language with differential object marking: when the linguist tried to elicit the 2SG-1SG mark for the verb for ‘eat’, at first the consultant said there were no such forms, until the two of themcame up with the idea of a children’s story with a man-eating monster in it, and then “You ate me yesterday” became perfectly plausible and the “correct” form of the affix emerged. Sorry for forgetting the details.
There’s also the question of cases where one construction is much commoner than another (entirely synonymous and structurally similar) construction but the rarer construction is still perfectly grammatical.
In Kusaal, “I think that he’s fallen down” can be either
M tɛn’ɛs ye o lu teŋin.
I think that he fall down
or
M tɛn’ɛs ka o lu teŋin.
I think and he fall down.
(where “and” is just a handy label for ka, which only sometimes means “and.”)
The Kusaal Bible has 219 instances of tɛn’ɛs ye for “think that” compared with a measly 31 of tɛn’ɛs ka, but as far as I can make out there is no difference of meaning at all, and my chief consultant actually always used tɛn’ɛs ka.
On the other hand, for “I see that he’s fallen down”, you can only say
M nyɛ ka o lu teŋin.
It was a long time before I realised that this is actually because verbs like “see” in fact take a quite different construction from verbs like “think” – a different construction which nevertheless can look exactly the same …
Assignments where Russian learners need to choose between perfeсtive and imprective verbs are sadistic:(
“I love Mary” or “I love Sue”, choose the right name.
Would the ladies in question not be offended by the use of the perfective?
Hanc amo, te amavi …
https://en.m.wiktionary.org/wiki/%D0%B2%D0%BE%D0%B7%D0%BB%D1%8E%D0%B1%D0%B8%D1%82%D1%8C#Russian
When in doubt, use vozljubit’!
Why? Especially if they don’t mind to be called “lady”.
Я вас любил…
Last night, ah, yesternight, betwixt her lips and mine
There fell thy shadow, Cynara! thy breath was shed
Upon my soul between the kisses and the wine;
And I was desolate and sick of an old passion,
Yea, I was desolate and bowed my head:
I have been faithful to thee, Cynara! in my fashion.
… [Cynara]
@PP:
Ah. Present perfect, as it were: “I have fallen in love with you.”
To be sharply distinguished from “I fell in love with you” (but then I discovered …)
@de
Undying love…
https://m.youtube.com/watch?v=8_8AyRodrI0
Now, someone, please, explain what the heck “Love as an act is missing (a? the?) verb” means.
Femina praeferri potuit tibi nulla, Lycori:
praeferri Glycerae femina nulla potest.
Haec erit hoc quod tu: tu non potes esse quod haec est.
Tempora quid faciunt! Hanc volo, te volui.
A truly Proustian message.
Past perfect “I had fallen in love with you” is a sentence unlikely to be spoken..It requires that you are no longer in love and it’s no longer a sensitive issue, but that at the same timeyou know the former love-object well enough to share this information. I can imagine two people meeting by chance and reminiscing many decades after the fact.
Actually if anyone would say that it would be Leonard Cohen. He was always falling in love, and then out of love, and then writing rather detached songs about the experience, sometimes addressed to his former love objects.
A truly Proustian message
It’s all there in his epic poem De temporibus amissis quaerendis …
Pullum says something (rather neatly) about the putative “infinity” of natural languages which never really occurred to me:
In other words, the “discovery” that “natural languages are infinite”, which would indeed seem to necessitate something like the generative method, is in reality not a discovery but a side-effect – an artefact – of the generative method itself and its underlying assumptions about what languages actually are.
It’s one of those things that’s obvious once it’s been pointed out …
I must point out, however, that “no amount of evidence” is wrong. An infinite amount of evidence would surely suffice to prove the claim (so long as it was judicially selected, of course.)
As it happens, I have a truly infinite amount of evidence that this comment box is too narrow to contain.
Wasn’t Chomsky’s proof of the infinitude of language the fact that, given any sentence, it is possible to produce another sentence referencing that sentence or including it? His example I was was sort of stupid but the general point seemed obvious. It would seem to be the default conclusion which someone would have to refute by proposing some theorietical moment when all possible sentences had been produced, but that’s ridiculous.
Pullum’s haiku example is idiotic. Haiku are an incredibly constrained form .
Perhaps I missed something but what he said makes no sense to me.
Wasn’t Chomsky’s proof of the infinitude of language the fact that, given any sentence, it is possible to produce another sentence referencing that sentence or including it?
Yes, but “it is possible to produce” actually assumes the conclusion. It presupposes (a) some sort of generative process and (b) that English actually itself is the infinite set you can create with such processes.
Pullum’s Haiku example is just to show that a feeling of open-endedness doesn’t actually prove anything: intuitions can err. Works for me …
It seems to me that the assertion of the default is what’s his happening here, and there’s less reason to assume that there might not be a generative process than there is to assume that there is one. (I don’t get “b”. Does it mean that even if there is generative process, at some point what is generated stops being English)?. So I say, it’s up to Pullum to prove that English is not infinite. Chomsky’s proof resembles a lot of math proofs,
Indeed, intuitions can err, but what does that have to do with this case?
Chomsky’s proof resembles a lot of math proofs,
And that’s exactly the problem: languages aren’t math, no matter how much Chomsky wants to pretend they are.
I mean, do you seriously think a speaker of English could produce a sentence containing a billion embedded clauses? And a billion, compared to infinity, is essentially zero.
I don’t get “b”
I mean, that he takes it for granted that the actual object of study is “all possible English utterances”, where “possible” is assumed from the outset to mean “which can be constructed by the very process I am about to propose.” It’s circular.
There is no such set, except as an imaginary production of the generative process itself. It is illegitimate to say that this figment is the target of linguistic enquiry: it is not, and no actual study of grammar could ever proceed in such a way. Instead, grammars are theories of how people make and understand utterances. Like other theories, they live or die by the successful predictions they make or fail to make. No genuine grammar is ever finished. The Chomskyan idea that progress in grammar reflects ever more refined application of a given theory, confirming its correctness over and over by ever more ingenious analyses, is a travesty of scientific method.
The Chomskyan approach is like stating that biology is the study of all conceivable possible life forms in the history of the universe. Biology is, of course, infinite … living things can always give rise to more living things, and there is no intrinsic limit to the process …
In fact, the latter-day Chomskyan approach to language, if applied to biology, would reduce that science to witterings about the élan vital (the rest is – ahem – butterfly collecting …)
And the problem is that “infinite” means something very different, and with different consequences, from very very very big. What is, let’s say, the total number of different compositions for a standard piano playable-in-principle by a human pianist (which limits how many keys can be pressed at once, how distant keys from each other pressed simultaneously can be, and how quickly the fingers can do something different in such-and-such way than they were just doing) that are no more than X minutes (or Y quarter-notes, if you prefer) long and meet such-and-such “grammatical” criteria distinguishing between “acceptable” harmony and “unacceptable” dissonance. Millions, billions, trillions, but not an infinite number. That’s a https://en.wikipedia.org/wiki/Fermi_problem, as is estimating the order of magnitude of the number of different possible (where “possible” has some real-world constraints built into it) English utterances
There’s that joke about how one can create formally grammatical sentences just by repeating “buffalo” an arbitrary number of times. However, even I, who am on record that I would not adjudge a sentence ungrammatical for being too recursively complicated to be comprehensible, find myself doubting whether that’s really English. For more realistic heavily embedded utterances, you have a sort of sorites problem; even if you are certain that a sufficiently deep degree of embedding is unacceptable, drawing a bright line for “how much is too much” is going to be very difficult. I would also expect the location of the line to have very low intersubjective consistency (something already seen in the linked thread); our parsers are trained almost entirely on language involving fairly low degrees of recursion, and what happens at very high degrees is essentially undefined behavior. (Okay, know that’s actually the wrong computer science analogy, but I can’t think of the right one. The tachycardia has passed, but I still feel a little weird.)
Indeed, there is nothing specific to language among human behaviours when it comes to this spurious argument that, as an object of study, it is “infinite.” I can tap my finger on this desk. I can do it again. And again … in fact, in principle there is no limit to how often I can do it …
[Spooky! But of course, it is because I have an innate finger-tapping organ in my brain …]
@David E.: But even then: a) it is overwhelmingly likely absent unknowable future changes in medical science that you will die before the age of 120; b) the number of seconds between now and your potential 120th birthday is finite and calculable; c) the average number of sustained taps-per-second based on your own prior observed finger-tapping and that of other members of your species has a finite upper bound that can plausibly be estimated; d) mathematical allowance can be made for minimalist assumptions about occasional periods of sleep etc. … So again potential maximum number of future finger-taps is another Fermi question with a fuzzy but finite answer where you can at least come up with a defensible order-of-magnitude guess.
First, it’s not one sentence that can contain a billion embedded clauses. It’s that any given sentence could have an additional embedded clause. (in writing at least).
And if language is not like math, why the hell are we arguing about the difference between an infinite number and a very large number, e.g., the number of atoms in the universe? Perhaps we should say “indefinitely large”.
It’s that any given sentence could have an additional embedded clause. (in writing at least).
And how do you get from there to “infinite”? I’m having a hard time understanding why you don’t see the absurdity of the argument.
If any given sentence can have an additional clause, then you have an infinite number of sentences, in the same way that you have an infinite number of numbers because you can always add one. As far as I know this si how numerical infinity as reached.
I am neither a linguist nor a mathematician and from time to time I’m glad. I really feel that this is the wrong place to attack Chomsky, though. It reminds me of the arguments over an open future vs. the deterministic blacvk universe and the argument whether the path evolution took was the only one possible or whether other outcomes were possible.
“She had swallowed the cow to catch the goat which she had swallowed to catch the dog which she had swallowed to catch the cat which she had swallowed to catch the bird which she had swallowed to catch the spider that wriggled and jiggled and tickled inside her and which she had swallowed to catch the fly which she had swallowed for reasons unknown to me, all of which makes me concerned about her near-term survival chances.”
What I take to be Pullum’s claim (which is interesting but perhaps could be rebutted) is that a literally infinite set of potential sentences makes it much easier to postulate the need for a generative machine that can generate it as output if you keep turning the crank indefinitely, whereas a quite-large-but-finite set of plausible-in-the-real-world sentences of a given actually-occurring natural language might be more plausibly described and analyzed via some other conceptual approach looking for patterns and regularities and whatnot that does not seem to require such a machine to generate it. I’d be interested in hearing a good rebuttal of Pullum’s point, but I must say I personally recall hearing the same infinite-set claim at or before my enrollment in Ling 110a 38 years ago this coming September, finding it superficially plausible, and never managed before seeing the link on this thread to think up Pullum’s point about how it doesn’t hold together.
I’m certainly open to a no-harm no-foul analysis that would argue that while some Chomskyoidanistites did end up peddling a false tale about infinity and trying to spin it as another reason their approach was right, they can afford to give that up without seriously undermining their remaining arguments, such as they are, because it was always sort of a gratuitous/nonce add-on to the justifications for the approach.
A separate question for those who know the history of the Former Times (B.C., or perhaps B.Ch., we might call them). Before the rise of Chomsky et al., how popular was the approach of setting out grammatical sentences next to ungrammatical sentences that are only slightly different in some specific way in order to illustrate a subtle rule. I’m thinking for example of this set, which is one of the few things I remember from Syntax I in fall ’85 (in which I received a very bad grade that I can’t say I didn’t deserve):
a. He ran up a big bill.
b. He ran up a big hill.
c. He ran a big bill up.
d. *He ran a big hill up.
I certainly don’t think you *need* a generative approach to explain why d is different from the others, but were other schools of thought (Jesperson? Bloomfield? Prague Circle dudes?) actually doing that sort of approach in an earlier era? Certainly the “here are our Horrible Examples of Obviously Wrong Sentences” bits you get in traditional prescriptive grammar books in the genre tradition going back to the 18th century have, at least for me, a different feel or vibe than this, and I still find it a useful pedagogical/analytical approach abstracted from the Chomskyan baggage with which it was associated when I first encountered it.
I don’t myself think there is anything radically wrong with a bit of transformation or even generation in grammars if it helps. As an expository technique, it can work fine, indeed can be much the most lucid way of presenting your description.
Similarly, the labelling of ungrammatical sentences as ungrammatical and discussion of exactly what it is that makes them ungrammatical in context can be helpful, and thinking about just what makes them ungrammatical can readily lead to actual linguistic discoveries.
The radical problem with Chomskyanism in all its protean manifestations over the years has not really been with any of this as such: it’s the incoherent philosophy behind it coupled with a disdain for proper scientific methodology (and the ludicrous claims about the Fundamental Truth the system encapulates, based on these two mortal sins.)
I’ve no problem with learning from Chomskyan investigations of specific linguistic phenomena (though disappointed that they seem to have much less to teach me than their rhetoric would imply.) Fas est et ab hoste doceri …
“(14) colorleaa green ideaa sleep furiously
which is a grammatical sentence, even though It is fair to assume that no pair of ita words may ever have occurred together in the past”
(I am not fixing errors in case someone will want to google it). The first counter-example:
“allowing the excess of alkali to evaporate, and adding a small portion of colourless green arabic”
colourless green arabic is unexpected.
(I am not fixing errors in case someone will want to google it)
There are countless search results for “colorleaa green ideaa sleep furiously”, all relevant. Might as well have searched for the original.
Google gets up with the cows.
[Is that last sentence “unexpected”? Whether or not you would say it is, what difference would that make to anything ?]
I find the word “unexpected” in this context to be unexpected. “Surprising” would make more sense to me, for reasons on which I may elaborate in a later comment here.
If the argument for infinity is not a single sentence, but a discourse where each utterance references the previous: Forum threads do not have to end because the OP is dead. There are threads in the Hattery that span more than 50 percent of the lifetime of the universe, I mean WWW. Some of them have seemed infinite indeed. (DAVID FOSTER WALLACE DEMOLISHED has been an adult for more than a year, it turns twenty in a few weeks).
The heat death of the universe will be marked by the last node in the Language Hat server net closing down, after posting its goodbyes in that thread.
‘There are countless search results for “colorleaa green ideaa sleep furiously”, all relevant. Might as well have searched for the original.‘
Showing results for “colorless green ideas sleep furiously”,
Search instead for “colorleaa green ideaa sleep furiously”,
If you click it, it shows only 2 results…
Nathaniel Hawthorne, The Blithedale romance (1852): “‘What a rich tinge it gives to his colourless ideas, when he speaks of Zenobia !’ I whispered.”
“… chasing crude thought slike half grown fledgelings into the brambles, or pulling up green ideas by the roots…“
I don’t understand why potential infinite recursion is supposed to prove anything about an innate language faculty?
If that is what is being supposed, then I guess it is based on an assumption that “potential infinite recursion” is not something that cavemen could have thought up on their lonesome.
So since speech production requires potential infinite recursion (a further assumption), the latter must be “innate”. In fact innate in the form of a language faculty – what else would it be needed for, if not for speaking ? Not for chasing bisons.
Note the infinite recursive dependency of these assumptions on each other !!
So far as I have read, Cantor didn’t invoke an innate language faculty to justify Aleph. He had an extensive correspondence on philosophical and theological matters relating to the notion of “infinity”, but none about parse trees or UG.
http://www.lel.ed.ac.uk/~gpullum/bcscholz/Infinitude.pdf
The paper by Pullum & Scholz that he refers to.
I don’t think that reviving 19c or whatever discussions of actual vs. potential infinity can help in understanding linguistics. For my taste, Pullum’s article is too heavy on philosophy. If you have to reach philosophy to explain real world phenomena, you are already on the wrong path (From Dovlatov. Nephew: Uncle, lend me some money, or I can end up on the wrong path; Uncle: you’ve already made your first mistake, you asked me for money). In addition, he tries to show off his math chops and [somebody’s name]’s law strikes back, in Euler’s formula there are no ternary functions and you cannot pretend that exponential function is on the domain of real numbers.
Yudkowsky thinks that in practice people can handle only 3 levels of recursion. Which probaly is just as meaningful as the “law” that almost all cases fall within 3 standard deviations of the mean. Rather heavily depends on the meaning of “almost all”.
Pullum and Scholz make various points, notably that the doctrine that “languages are infinite” is (potentially) true only if you actually identify a language with the set of all possible grammatical utterances in that language. There is no need to do this if your model of what a language actually is comes from a constraint-based grammar; but in fact there is no need to reify the set generated by generative rules even if you are a generativist, much less to identify this set with the actual object of grammatical study. Chomsky himself has abandoned this conception of language, which is a hangover from a specifically mathematical definition of “language” prominent in his early work.
Moreover, generative frameworks need not produce infinite sets anyway (they give a counterexample.) In other words, the doctine is not only unmotivated, but also has no consequences: it is irrelevant to any substantive point at issue.
P and S go on to speculate on the reasons why this chimaera is routinely presented as a “fact” about “language”, suggesting inertia (not keeping up with ANC’s latest aperçus) or imagining (wrongly) that this “fact” is itself evidence for some part of some Chomskyan schema, or that it dishes the behaviourists.
If you have to reach philosophy to explain real world phenomena, you are already on the wrong path
Surely, that is itself a philosophical position?
I’d say that (like theory) philosophy is in fact inescapable. If you think it is irrelevant, it means that you are either unaware of the preconceptions through which you view the world, or believe that they can never mislead you on any important matter. I myself lack such a firm faith in the inerrancy of my preconceptions …
I think P and S go heavy on the philosophy because the fundamental problems with Chomskyanism actually boil down to (very bad) philosophy. (Esa Itkonen is all over this point.)
Search instead for “colorleaa green ideaa sleep furiously”
I suspect Stu didn’t bother with the quotes, which of course vitiates the whole exercise.
LH, without quotemarks I still obtain 2 links, the first one to the pdf where I copied it from (tiikee models for tie description of language…)
“Chomsky himself has abandoned this conception of language”
@DE, but they also say early Chomsky did not claim this.
—–
Omer above suggested to distinguish “grammaticality” (theoretical) and “acceptability” (experimental). Acceptability depends on what speakers call “acceptable”.
It makes sense, and it is a good idea to define what we mean by grammaticality first.
Same about “sentence”.
… or “expression”: “The collection of all grammatical English expressions is an infinite set.”
—
Pullum & Scholz cite literature suggesting a complete lack of recursive hypotaxis and syndetic coordination in various languages: early Akkadian, Dyirbal, Hixkaryana, Pirahã, Proto-Uralic, Wargamay, and probably many others.
From the paper by Pullum cited earlier here. Proto-Uralic?:/
I think that it’s a different question, whether particular individual languages have the property that an orthodox generative grammar of one of them would generate an infinite set of acceptable utterances.
I think most of the interest in this intrinsically dull question has come from a belief that the demonstration that such languages exist is either necessary or sufficient to disprove “generativism”; I agree with both Pullum and the Chomskyites that this is not in fact the case.
Recursion/embedding seems more important in this context than coordination. However, I’m not myself convinced that this distinction itself is quite as robust, fundamentally, as is sometimes supposed. For example, in many West African languages (Kusaal, inevitably, among them) narrative, of the kind that could go on potentially for as long as the narrator has not bored her audience, features sequences of clauses which are either very like successive subordinate clauses or actually formally indistinguishable from subordinate clauses; conversely, it’s not hard to find languages in which the equivalent of SAE relative clauses are formally paratactic. (The – I thought, rather convincing – orthodox Chomskyite riposte to Everett’s claims about Pirahã attempts to show that Pirahã does too have recursion, just not recursion nicely marked comme il faut.)
Is “This is the woman that swallowed the X, that swallowed the Y, that swallowed the Z …” really a profoundly different sequence cross-linguistically and sub specie aeternitatis from “This is the woman. She’s the one that swallowed the X. And that’s the one swallowed the Y. And that’s the one swallowed the Z …”?
One reason “Pullum & Scholz” as well as subsequent Pullum-only writings that cite P&S might be rather philosophy-heavy for the tastes of some is that Scholz (who was prior to her comparatively untimely death Pullum’s wife as well as co-author) was not a linguistics person as such but an academic philosopher by both training and job title (albeit with a master’s in “cognitive science” thrown in along the way). Whether her personal relationship with Pullum led her to focus on “philosophy of linguistics” as a specific subgenre of “philosophy of science” or whether a prior interest in that led to the personal relationship is not known to me.
That said, her dissertation (completed 4 years prior to her marriage to Pullum), was on the linguistics-adjacent side of academic philosophy. The abstract:
In Wittgenstein on Rules and Private Language, Saul Kripke developes a forceful and elegant interpretation of a skeptical paradox found in Ludwig Wittgenstein’s Philosophical Investigations. If cogent, the paradox shows that a broad class of syntactic, semantic, and psychological theories rest on a mistake. The mistaken idea is that there are facts which uniquely determine the semantic properties of syntactically individuated expressions in advance of their considered use. The paradox has epistemological, as well as ontological, import. The facts in question are alleged to do double duty: they both constitute the semantic type of words and justify meaning and reference attributions. ;In the philosophical literature it was alleged that causal theories of reference provided a solution to the paradox. I marshall arguments to show that no purely causal theory of reference can resolve either the ontological or the epistemological form of the paradox. I also show how a modified causal theory of reference that appeals to the evolutionary history of perceptual mechanisms can provide an adequate solution. ;The solution that I articulate has limited scope in two senses. First, it applies to a relatively small class of count nouns and proper names where the denotata of the word is perceptually detected by means of shape. Secondly, my solution does not address the epistemological version of the paradox. Since my solution is unlikely to generalize to expressions with abstract denotata or to substance terms, and since the reader is owed an account of the grounds of reference attributions, I take up the evaluation of so-called “skeptical” solutions to the paradox. I develop a taxonomy of conceptually distinct skeptical solutions–i.e., alternatives to both skepticism and traditional theories. A synthesis of several distinct but compatible views provides the most attractive skeptical response to the paradox
FWIW, one well known children’s song ends with this stanza:
Hear it here. This is not embedding, strictly speaking, but it’s equivalent to “the boy that lay on the pillow that came from the feather that was on the bird that came from the egg that lay in the nest that was on the leaf that sat on the twig that sat on the branch that sat on the tree that stood on the hill that lay far away in the forest.
But even though that is not “nested” either, you might attach importance to the fact that tradition did not chose that construction. Or not.
you might attach importance to the fact that tradition did not chose that construction
It’s all part of the Scandi-Congo Erbgut.
However, the Scandinavian branch has abandoned the original formally subordinate narrative clauses. I blame the Sagas.
@Lars: As I now realize may not have been obvious to those who did not have an anglophone childhood, I edited/rewrote the penultimate (in some versions) stanza of a well-known children’s song into a formally embedded/nested structure that is not present in the conventional version of the lyrics … So presumably the same thing where you might or might not attach importance to the fact that tradition did not choose that construction.
“Is ….really a profoundly different sequence cross-linguistically and sub specie aeternitatis from …”
I have no reason to think so (or otherwise).
“formally” ideally should include intonation.
And then there can be other hints (excluded from langauge)…
‘They proceed to the rather surprising claim that “The core property of discrete infinity is intuitively familiar to every language user” (we doubt this)‘
I think I am with ‘them’ here… (partly, because every is too much). I mean, yes, as a child I was aware of what is “and”. I mean, I was aware that I can say “I love Mary and Sue and Mary and Sue and Mary…” and keep saying it until I forget who is the object here. “…and Mary and Sue love me” (I am monogamous, it is just my example).
Let’s also do not forget that numerals are a part of language too.
But you can’t keep saying it forever. The principle is obvious; the extension to infinity is nonsense.
@drasvi, well with numbers you also hit pragmatic constraints, such as the demonstrations that writing out a “googolplex” in standard decimal notation in a standard font size would require more sheets of paper than the mass of our galaxy and would require more time to say out loud than the entire elapsed time from the Big Bang until now.
I wonder if in the math-nerd subculture there is a known record (which people with time on their hands sometimes try to break) for “most digits of the decimal expansion of π recited out loud by a single human being in a single session”? (Let’s assume the contestant is allowed to read from something rather than reciting from memory, which would be a tighter constraint.)
@JWB:
Unfortunately I can’t access Barbara Scholz’ PhD thesis (apparently she wasn’t very happy with it and declined to have it published as a book), but that actually makes a lot of sense in this very context. In fact, I was just thinking of St Ludwig’s views on rule-following, which bear greatly on this whole question of “infinity” – in mathematics, let alone linguistics.
Searching “Kripke-Wittgenstein Paradox” turns up lots of papers on this.
It’s not by any means clear that Kripke’s Wittgenstein is the One True Wittgenstein.
@LH, I was making a different point: that you don’t have to be a linguist to notice what “and” can do (and other funny things about langauge). I do not know what part of the claim they find “rather surprizing” but if they mean that most people never think about such things, likely they are wrong.
@JWB, yes. It just occured to me that there is no reason to limit ourselves with “clauses”.
I thought WP must have something about it somewhere …
https://en.m.wikipedia.org/wiki/Wittgenstein_on_Rules_and_Private_Language#The_rule-following_paradox
And speaking about this: I do not think, for example that formal languages are unrelated to natural languages. How? Formal languages are used by users of natural languages, likely because they are users of natural languages:) By shamans for their logical rituals.
Similarly, sequences of symbols can’t be equated with natural languages and possibly speech can be represented on paper in other ways, but so are our scripts and we can read aloud what we wrote. And then we devise other sequences and find ways to read them out loud too.
Formal languages are an extension of natural languges. If someone is interested in what human mind can do with language, they have to do with the quesion. Human mind can do things with them.
Surely, that is itself a philosophical position?
It was an observation from a number of books on philosophy of science that I’ve read a number of decades ago. It is useful to judiciously drop one’s preconceptions at an appropriate time, but philosophy is no help in figuring out which preconceptions and at what time. If someone has something useful to say about language, they should say it without regard to with which school of medieval scholasticism it most accords.
It is useful to judiciously drop one’s preconceptions at an appropriate time, but philosophy is no help in figuring out which preconceptions and at what time.
“Drop one’s preconceptions at an appropriate time” sounds like a game of strip poker. I can’t imagine why you would expect philosophy to help you in that.
If someone has something useful to say about language, they should say it without regard to with which school of medieval scholasticism it most accords.
You must have read pretty boring books on the philosophy of science, if all that has stuck is the dismissive label “medieval scholasticism”. You’re dismissive because the books were boring. It might be time to drop the preconception that boredom is a natural condition.
I must say that I am bored by people who use boredom as an excuse for having nothing interesting to say. But who cares whether I’m bored ? Who cares whether anyone else is bored ? Even strip poker is more exciting.
As someone said: if something is not worth doing, it’s not worth doing well.
“Colourless green arabic” is plainly enough an error of some sort for “colourless gum arabic.”
It might be a complaint that the style of typical Arabic works on environmentalism is not very poetic.
“Colorless Green Stamps” because all the ink bled out when they fell into the bathwater.
Rodger C: well spotted. It’s not an OCR error, the original does say “colourless green arabic”, but it’s obviously a garbled copy of an earlier publication: “The most beautiful of all the red inks, is made by a solution of carmine in liquid ammonia, allowing the excess of the alkali to evaporate, and adding a small portion of colourless gum arabic.”
Rodger C, ktschwarz, aha!
I was not sure, because “gum arabic” is a well-known name, but it also was called “arabic”.
school of medieval scholasticism
Don’t diss the scholastics. They were clever buggers …
I feel about Nominalism (not a single thing, in fact , as Pullum and Scholz rightly point out) rather the way that some adolescents apparently feel about labels for different forms of sexuality: Hey, I’m a nominalist! There’s an actual name for it! It’s not just me. And some famous people have been nominalists …
[No ideas but in things …]
Their pouint about haiku is different by the way:
“To see that infinitude is not necessary (and here we are endorsing a point made rather differently in the philosophical literature by Gareth Evans 1981), it is enough to notice that creating a verse in the very tightly limited Japanese haiku form (which can be done in any language) involves creation within a strictly finite domain, but is highly creative nonetheless,“
I found where Chomsky writes about it! “The Galilean Challenge: Architecture and Evolution of Language”
https://iopscience.iop.org/article/10.1088/1742-6596/880/1/012015/pdf
He names Galileo, Arnauld and Lancelot, Descartes, von Humboldt and Turing, all in connection with the very same idea. But it is a lecture for physicists….
“I thought WP must have something about it somewhere …”
I did not, but
https://en.wikipedia.org/wiki/Digital_infinity
But if you need an infinite sequence that does not repeat itself, you can toss a coin…
there are a lot of exceptions, like “The dog is a noble creature”.
I’ve never seen anyone point out that “The dog is a noble creature” is a different (mostly written) register.
“Dogs are dirty animals” sounds natural in speech where “The dog is a dirty animal” does not (except as a deliberate elevation to a different tone). You wouldn’t want to teach “The dog is a dirty animal” to people who are mainly want to learn how to communicate in colloquial English. For those who wish to learn how to write English, however, it is a useful construction.
WP:
An axe is an implement…
The axe is an example…
Sigh.
I’ve been extremely busy recently and it’s a pity Omer disappeared before I could engage.
I’d just like to ask about these examples (concocted):
To eat without chewing.
To eat something without chewing.
To eat something without chewing it.
Are these acceptable or unacceptable? Do they have different meanings / usage situations? Do they relate to the issue raised?
* I know they’re not great examples but they’re the best I can think of at the moment.
And, of course, ‘eat’ is one of those verbs that don’t require an object.
Phillips referenced by Sprouse was comparing:
The outspoken environmentalist worked to investigate what the local campaign to preserve the important habitats had harmed __.
The outspoken environmentalist worked to investigate what the local campaign to preserve __ had harmed __.
http://www.colinphillips.net/wp-content/uploads/2014/08/phillips2010_armchairlinguistics.pdf
@drasvi
In that example, I would say there is no gap after harmed (the word “what” means “the thing or things that”, as in “You know what you did/are doing.”)
The outspoken environmentalist worked to investigate what the local campaign to preserve the important habitats had harmed this time.
The outspoken environmentalist worked to investigate what the local campaign to preserve fruit had harmed with their gigantic gooseberry harvesting machines.
@PP, I think the idea is that the object of “harm” is “what”, thus “harm” is naked, and naked “harm” allows “protect” to be naked too.
Actually supplying it with object is not easy.
“The outspoken environmentalist worked to investigate what the local campaign to preserve the important habitats had harmed __.,”
does not imply for me that the campaign harmed habitats. I can read it in two ways:
1. The OE does not know what it harmed. She wants to find out (they MUST have harmed at least something…)
2. The OE knows what it harmed. She wants to study the harmed thing more closely.
The outspoken environmentalist worked to investigate what the local campaign to preserve __ had harmed __.
This makes me think that the campaign harmed the very same thing it was [a campaign] to preserve.
Then, again,
3. The OE does not know what it is
4. She does, and is interested in that thing (but not in the harm).
If any of this is what is meant, then “what the camplaign to protect what it was going to harm…” would indeed be quite complicated.
In all readings the OE is not interested it the slightest in what less experienced and less outspoken environmentalists are usually interested in (nature and amount of damage and ways to fix it).
Stu, I actually mostly read about positivism. And I liked it. It’s just not helpful in understanding something about the real world, but it is a fun investigation of how people can think about their thinking. As for medieval scholasticism, it’s not me it’s what Pullum (and apparently Katz) are interested in.
Logical positivism dealt with all problems of any philosophical interest by declaring that they didn’t exist. It’s not helpful for understanding anything. (Especially not the philosophy of science, which it characterises in a wholly impossible way, incompatible with any actual real-world scientific methodology.) Philosophy for the perpetual grumpy teenager in all of us …*
I’d say that much the most important task facing the philosophy of science is separating real science from pseudoscience. Far from being an irrelevant ivory-tower pursuit, this is surely more important now than ever. It’s also by no means a trivial question with obvious answers.
* As Ayer said in his introduction to a reissue of Language, Truth and Logic, “It is in every sense a young man’s book.” (Enjoyable, though.)
But is there a need to do it?
If someone came up with an excellent idea but can’t explain anything, it is real science. There is no way to tell it from pseudo-science other than spending a lot of time on reformulating and experimenting.
I mean, this someone suggested something based on intuition or this someone published a theory formulated in a crazy langauge.
If you trust her intuition – you do it, and then you find out that the suggestion was worthy.
Like drasvi I don’t think that separation of science and pseudoscience is really possible. I mean, people should observe some sort of scientific hygiene like not ignoring contrary evidence, being open to discussion etc., but there is no guarantee that any particular idea will prove fruitful or completely wacky. If people are trying to communicate with the souls of the departed, just let them. As long as they are not moving the table with some contraption.
But is there a need to do it?
Oh, YES. Yes, indeed.
someone suggested something based on intuition
No problem. Scientific theories can’t be based on much else. Certainly not deduction.
and then you find out that the suggestion was worthy
How? (This is the whole crux of the matter.)
people should observe some sort of scientific hygiene like not ignoring contrary evidence
Why? Why would you do that?
And what counts as “contrary evidence”? Where do you find it? How do you find it? Why bother?
there is no guarantee that any particular idea will prove fruitful or completely wacky
Of course not. How could there be?
So, how do you tell if an idea is wacky?
What do you do next if you decide that it is? Why?
Pseudoscience kills people.
Sure, it’s not a trivial matter to distinguish it from real science.
Happily, that does not mean that nothing can be done to separate them at all.
Real science … That’s merely the plot of a zillion sci-fi movies. A crazy idea is put into practice because nobody knows a better way to escape the aliens.
They have only that one shot, so it’s not an experiment since non-reproducible. That the protagonists then survive is not a result worthy of science, but of the storyboard.
In der Not frißt der Teufel Fliegen.
(Lord of the Flies), Aramaic ba’al-debaba.
A crazy idea is accepted because nobody has a better plan to escape the aliens.
Absolutely. And if it works, you can see if it works on the next bunch of aliens too.
Apropos of SF and science:
https://www.myconfinedspace.com/2010/03/10/most-mad-scientists-are-actually-just-mad-engineers/
There sure are a lot of woke hippie musings on science being floated here. I suspect the piss is being taken. As the Duchess, cook and baby put it: “Wow! wow! wow!”
Everyone should go off and read Against Method until they are sorry.
(Also, Feyerabend is a lot more fun than Ayer.)
True, but Ayer was good enough for me at age 16 or so. He was polish to the silver spoons of my mentality at the time. I had not yet come into contact with Advanced Thinking.
Now I pensive dance with the daffodils, lonely as a cloud.
Well this thread, from which I have resigned, bids fair to continually approach infinity without ever (for some reason) actually attaining it. Pseudoscience: Newton spent the majority of his scientific life studying alchemy.
(I now retreat back behind the curtain).
You can resign, but you can never leave…
There can never be a final comment in this thread, because it is always possible to add another.
It follows, therefore, that this thread is countably infinite …
…
…
…
Well, nobody has claimed more than that for language innit?
Actually, yes they have …
https://languagehat.com/chomskys-forever-war/#comment-4386589
The universal personality (UP) generates infinite number of individual commenters.
If we some day manage to teach computers to be actually “speakers of English” (neural networks are currently doing an increasingly good job at at least the superficial aspects of it), it seems entirely possible that one that is exceptionally well-endowed with processing memory could furnish a billion-deep embedded clause… for that matter, we already possess the means to write down such sentences, even if to be left unparsed in its whole by any living mind so far.
I am not a fan either of the idea that language “already is” infinite; but as long as we equip it with some modest amount of rules, a sense of potential infinity arises very, very easily. I mean I would be curious about a Radically Ultrafinitist Ruleless theory of language too, but the feasibility of that is no foregone conclusion either.
A much more assailable stopgap of the “infinitarian” approach is in it forgetting the fact that humans — the actual users of Englishes and other such constructs — labor under memory (and processing) constraints, hence always limited to some finite amount of sentences that they can parse within some finite amount of time. (Er, that we can parse. Was that UP speaking there for a moment?)
“How? (This is the whole crux of the matter.)”
I mean, it is real work. If you feel that there is something to it, you work on it, but you can’t know beforehand.
I am at the moment speaking about “science” as a way to learn more, curiousity. Here you risk wasting your time. There are sciences that make good predictions, you can read about them, but usually we do not count learning those as “science”. “Science” is exploring terra incognita, exploring new methods and finding new terras to explore. So is not it novel, in some ways unpredictable and creative? Why rules?
And you need to waste lots of effort to see that you have arrived to some place or that you can make predictions (“or” because imagine you invented Lobachevsky geometry), and if the idea was wrong you risk going in circles for years. A mathematician publishes a sketch of a proof and then another team works to complete it. Why do they do it? I think because (1) they trust the guy (2) the sketch does look like something that must work (intuitively).
Generativists are doing the same. Many students were forced to study generative syntax and feel that studying something else could be a better idea, but that is a different issue…
Making good predictions is easy if you get to choose what to predict.
Astrology often does that. So does generative grammar, come to that.
(Upping the ante) Newtonian physics makes good enough predictions to land a man safely on the moon. Yet it is now known for certain to be an incorrect theory.
What about making wrong predictions? How do you find out if your theory does that?
How do you react when it does? (Not a rhetorical or trick question. The answers to such questions are neither obvious nor simple nor settled.)
Or Einstein. Apparently Einstein had a blog (…or five seconds on the radio?) where he posted a Quote of the Day every day.
Oh yes, that works, actually.
” :-/ Ich weiß nicht. Ich weiß echt nicht[, ob wir das tun sollen].”
Oh.
I did; we were taught it, and it occurs in quidquid id est, timeo Danaos et dona ferentis. But I guessed the meaning of deperditus wrong, and the meter doesn’t make the vowel length of sensus clear, so I went for a simpler and wrong interpretation. I’m not sure if I was ever taught the accusative of respect.
Geschlafen sein, as opposed to the normal geschlafen haben, has made it into this category in *handwave* Austria. Possibly all the way north to the White-Sausage Equator, given that (as I recently learned) gesessen sein (standard in Austria) is used even by Franks.
I agree, but this doesn’t tell us anything universal either. German doesn’t have an adjective like asleep paired with the participle sleeping, but does have lebendig “alive” vs. lebend “living”; and ein lebendiger Hund is perfectly cromulent.
You might indeed become your own grandpa.
Quoted for truth.
David Marjanović: You might indeed become your own grandpa.
It sounds funny I know,
But it really is so.
Russian has this situation with я устал “I’m tired” where I can’t really tell whether it is an adjective describing my present state or perfective verb describing what I did (I blinked, I arrived in the city, I got tired, I fell in love with it).
It’s pretty common in West African languages for the same verb form which expresses perfective aspect in dynamic verbs, with a past or present-perfect sense, to be used with present meaning with stative verbs.
Biblical Hebrew does that too. And you have preterite-presents in the older Indo-European languages based on the perfect form: I believe that there are theories that the stative sense of the perfect form was originally primary.
Thanks, I was too lazy to look for it. 🙂
Like, uh, the modern Germanic languages, where they’re still inflected like preterites ( < perfects) and the other tenses are just made up: will, can, may, shall… also weiß.
That’s mainstream for PIE: “I’m in the state of seeing” > “I’m in the state of having seen” > “I know”.
@DE, yes, it occured to me that this mystery can be a key to another conspiracy that gave rise to a large international network of tense-aspect-mood.
I think that the Bioprogram enthusiasts (if there are any left) think that creoles naturally develop the sort of system I describe as common in West Africa; certainly the English-lexifier Atlantic creoles have it, but it seems to me that there is an obvious explanation for that which is a lot more plausible than attributing it to some sort of universal tendency. I don’t think Tok Pisin and her sisters have a system like that (though Joel will know a lot more than me about that.)
A hypothesis that can be rejected or confirmed is a good thing…
Owens (because I wanted to check Arabic creoles and Russian pidgins and was lazy to consult Bickerton :
Again, I was too lazy, but the above does look like he is approximating an idealized situation:
(1) Mowgli-talk [ideal] (2) development from a simple seed, e.g. an instant pidgin can be used as the seed [idealized setting] (3) this thing.
By Mowgli-talk I mean what a population of Mowglis would have invented over a few generations.
(1) is a mad scientist’s dream
(2) may retain some features of 1 but one expect interference from pragmatics/semantics/structure of the seed (a pidgin).
(3) approximates 2 and can be studied. Thus (3) provides a glimpse into a mad scientist’s dream…
Tok Pisin is not really “early-creolozing creole” then…. But can still be informative as an approximation of mad scientist dreams, maybe. I do not know.
“creoles naturally develop the sort of system I describe as common in West Africa; ”
Imagine a pidgin where nothing is marked. It must look like “I read it”, “I read news”, “I big” and any distinctions only can be derived form contexts.
There are actants, but not statives-shmatives. You can speculate about what the speaker has in her head as you please. Such pidgins do not differ much in morphology, but they can differ in pragmatics, semantics and distribution of words. Do they?
Studying such differences by traditional linguistical means (publishing isolated lines and trying to observe what is not there, but what traditional grammar wants) is futile.
Now the claim about relatedness of stative and perfective would look as: “there are words that are usually used with reference to past punctual events and words that are usually used with reference to present” (presumably future, past habitual/continuous and present punctual are either not referenced at all or referenced differently).
Everything becomes easier when we have some sort of TAM markers.
So as I understand the claim is that one group of words obtains a marker when it refers to present, and all groups obtain another marker (“anterior”):
(now) I smart, I ajump
(yesterday) I been stupid, I been ajump, – iterative, habitual, continuous et.c
(yesterday) I jump – punctual event
(yesterday) I been jump – ???
Nubi seems to be vaguely similar….
…but not entirely absent, obviously. I haven’t looked anything up, but I’d expect the pidgins that early-creolising creoles formed from to have a grammar that starts more or less from the lowest common denominator of all the input languages. That would already explain why these creoles all have SO verb order (haven’t looked up if they all have SVO specifically).
Russian-Chinese trade pidgins are SOV and Russenorsk (where my nickname comes from) too.
That’s amazing. Occurring in none of the input languages, SOV must therefore come straight out of the bioprogr… oh.
Wiki sayz its “SVO, with some alterations for questions and sentences with adverbs” and “the tendency to move the verb to the final position when the sentence has adverbs”. The latter seems to be the normal Russian attitude “Телефон лежит на столе” changes to “Телефон на столе плохо лежит” (is ready to fall) is more natural then “Телефон плохо лежит на столе”, but you can find examples where SAdvVO is more natural, and in general Russian is not very particular about the word order.
Russenorsk supposedly had just two contributing languages*. It’s not unnatural if the Norwegians orverdid the SOV they heard in some sentences from Russians. We do tend to overdo the verb-final thing in German.
* I’m not so sure. For one thing, Dutch and English trade had been around the Finnmark coast for generations. I also remember seeing a suggestion that Finnish contributed both grammar and vocabulary. Many of the contributing speakers were coastal Sami and Kven (migrant northern Finns).
Russian-Chinese trade pidgins are SOV: so are almost all the languages spoken in the relevant area except Russian and Chinese; haven’t looked into the details, but I imagine that must have had something to do with it…