The Rise of Arabic.

In 2018 I waxed enthusiastic about Elias Muhanna’s New Yorker piece about Ahmad Al-Jallad, a professor of Arabic and Semitic linguistics at Leiden University, and in 2020 about Al-Jallad’s Twitter thread on the origin of the word Saracen; now I’m here to recommend his lecture The Rise of Arabic, posted on YouTube last December. I don’t normally link to, or even watch, hour-long videos — I like to absorb information via the written word — but in this case his presentation is so lively and informative, and the ability to see the inscriptions and maps (and hear his pronunciations) is so useful, that I wasn’t bored for a minute. He says the question of why there seems to be so little early Arabic, why it seems to appear suddenly out of nowhere, is mistaken; it conceives “Arabic” as “what is described in Western grammars,” including the definite article al- and a certain form of past tense, but that renders not only earlier forms of the language but modern dialects non-Arabic. If we talked about Egyptian the way we do about Arabic, we’d conclude the language of hieroglyphs was entirely different from that written in the Coptic alphabet and wouldn’t be able to see a developmental line connecting them (he also uses Old English as an example, at which point, around the 25:30 mark, we get Alex Foreman reading a sample). He says there was an explosion of Arabic scripts in the second half of the first millennium BCE, but the modern script begins with Nabataean (which was originally used not to write the users’ native Arabic but the official language, Nabataean Aramaic, so when it started being used to write Arabic there were many Aramaicisms, the longest-lasting of which was bar ‘son’). It’s an hour well spent if you have any interest in the topic.

Comments

  1. Sounds interesting.

    Sorry for revealing my ignorance, but I knew virtually zilch about Arabic till recently when I had to read about the development of Arabic linguistics. The history of those early days of both the Koran and the language (and the development of grammar so that it could be taught to non-Arabic speakers who had converted to Islam) is much more interesting and convoluted than I suspected. (E.g., the fact that many of the people who were intimately familiar with Mohammed and the Koran were killed in battle, making it harder to fix the definitive version.)

    One question I didn’t find an answer to was: Did the ancient Greek grammarians influence the development of Arabic grammars? That is, did grammarians like Sibawayhi come up with complete grammars of the language virtually out of thin air, or did they owe some kind of debt to (for instance) the Alexandrian grammarians?

  2. The history of those early days of both the Koran and the language …

    Al-Jallad’s claim seems to be that the ‘early days’ of Arabic were centuries before the Koran.

    the fact that many of the people who were intimately familiar with Mohammed and the Koran were killed in battle

    Ah. Yes, I must reveal a similar level of ignorance. And yet enough people survived to manage to have a schism immediately upon the death of the Prophet.

    If you’re aiming to found a world religion, might be better that there’s no commitments in writing as at the time of your death. Wait a few decades/centuries until you can ally with some temporal power base, _then_ write stuff down (in sufficiently reinterpretable terms) to align with them. That at least postpones the schisms until you’ve won more hearts and minds.

  3. wp on Sibawayhi quotes his biographer:

    He arrived in Baghdad, fell out with the local grammarians, was humiliated and went back to some town in Persia, …

    So there were already grammarians of Arabic. Sibawayhi was not a native speaker. (Compare Jespersen’s grammar of English.) Coming from Persia, it’s at least possible he was aware of Indian grammarians?

  4. Yes, he had an argument with a rival grammarian about the correctness of a form of Arabic. Sibawayhi maintained that a certain form was correct. The other grammarian maintained it wasn’t, and had bribed a couple of men to come in and back him up. That was the humiliation. The thing is, the other grammarian was also a Persian.

    In those days it appears that non-native speakers were instrumental in writing grammars of Arabic. Which makes sense, because it is often the foreigner who is better able explain grammar than the native speaker.

    I have no idea whether he was aware of either Greek or Indian grammarians. It would be interesting to know.

    PS: I haven’t had time to watch the video yet. I’m just writing from what I’ve read.

  5. J.W. Brewer says

    And/or it may be the foreigner who is better able to market his products to other foreigners interested in learning the language and its grammar (assuming the native speakers are not so nervously clueless as to think they are in need authoritative guidance, which will be a better assumption in some times/places than others)?

  6. And/or it may be the foreigner who is better able to market his products to other foreigners interested in learning the language and its grammar

    Not sure whether or not this applies to the Arabic situation.

    There was a religious aspect involved in all this. People who had converted to Islam were saying the words wrong, sometimes with religiously undesirable results. That’s why caliphs got involved. If I remember rightly, prejudice against non-Arabs getting the language wrong was countered by pointing out that they were all Islamic brothers.

  7. I’ve watched the lecture. Indeed very interesting.

    With regard to the antiquity of Arabic, I should point out that grammarians like Sibawayhi didn’t only use the Koran as a source of Arabic; they also used older songs and verses used by Arabic speakers.

    I do wish someone more knowledgeable than me would comment. I have only general impressions that I got from my reading.

  8. There was a booklet from 1977 by Versteegh, Greek elements in Arabic linguistic thinking. The title sounds promicing, but I haven’t read it yet (honestly, most of titles in Brill’s series Studies in Semitic Languages and Linguistics sound promicing:/). sci-hub.

  9. Personally I wonder what Persia herself looked like (in terms of linguistics)…

  10. David Eddyshaw says

    it is often the foreigner who is better able explain grammar than the native speaker

    True. With your own language, the very familiarity tends to blind you to things you take for granted which immediately strike an alert foreigner as strange and in great need of explanation.

    grammarians like Sibawayhi didn’t only use the Koran as a source of Arabic; they also used older songs and verses used by Arabic speakers

    And indeed the Qur’an itself has been retrofitted into the grammarians’ Classical Arabic: this is the reason for some of the quirks of Arabic spelling, for example, as with the writing of /ʔ/ and of the case endings.

    The lecture is very interesting, though the general thesis is largely familiar from Al-Jallad’s chapter in

    http://langsci-press.org/catalog/book/235

    (Other chapters are also fascinating.)
    His A Manual of the Historical Grammar of Arabic is here:

    https://www.academia.edu/38100372/Al_Jallad_A_Manual_of_the_Historical_Grammar_of_Arabic

    It covers the ground very thoroughly.

  11. I’ve now worked through the lecture; gosh it’s dense. As Hat says, seeing the inscriptions and maps is crucial.

    Al-Jallad’s main historical point is a language is not a script — by coincidence Prof Mair has a post with exactly this message wrt Sinitic languages on LLog as of a couple of days ago.

    So from the first half of the first Century BCE, there are inscriptions (chiefly prayers) in a tongue that is an early form of Arabic — but written in whatever script was to hand in the place (Southern Levant/NW Arabia) — and in which the writers named themselves as ‘Arabs’.

    Something identifiable as today’s familiar Arabic script didn’t appear until about the time of the Prophet.

    The preso shows a timeline at about 49:30 that summarises the evolution of the scripts.

  12. I do wish someone more knowledgeable than me would comment.

    Yes, where’s Lameen when we need him?

  13. One feature of both Arabic and Greek grammar which has been cited as a possible influence of the latter on the former is the use of words for “to hit” in paradigms: daraba, tupto.
    One would like to believe that this was an intellectual influence rather than a reflection of the word being prominent in people’s minds owing to corporal punishment in grammar classes…

  14. Bathrobe: In answer to your question on Greek influence (or lack thereof) upon the Arabic grammatical tradition, there is a good discussion of the issue in Esa Itkonen’s UNIVERSAL HISTORY OF LINGUISTICS. Drasvi: He also mentions in passing that the Indian grammatical tradition was known in pre-Islamic Persia. The book struck me as very thoroughly researched, and I think both it (a wonderful read) and more specifically its bibliography would be good starting points if you (or any hatter, really) wish to explore these and related matters more deeply.

  15. Universal History of Linguistics: India, China, Arabia, Europe:

    This wide-ranging book presents the linguistic achievements of four major cultures to readers presumably conversant with modern theoretical linguistics. The chapter on India discusses in detail Pāṇini’s (c. 400 B.C.) grammar Ast-adhy-ay-i as well as the work of his commentators Kātyāyana, Patanjali, and Bhartṛhari. In the Chinese tradition, the Confucian doctrine of the Rectification of Names’ is singled out for treatment. Arabic linguistics is represented by Sibawaihi’s (d. 793) grammar al-Kitāb, in particular its syntax, as well as the subsequent commentary tradition. The chapter on Europe, which is the most comprehensive of the four, covers the time span from antiquity to the 20th century; special attention is devoted to the contributions of Plato, Aristotle, the Stoics, Varro, Apollonius Dyscolus, and the Modistae. The achievements of the cultures in linguistics are treated throughout from a deliberately value-laden point of view. The achievements of Western antiquity and the Middle Ages are shown to be much more than the average linguist is inclined to believe. Even more importantly, it is shown that the Indian and the Arab traditions have been superior to the European tradition at least until the 20th century. The fact that a linguistic theory created some 2,400 years ago is fully as adequate as our best theories today must have far-reaching implications for the notion of ‘scientific progress’. More precisely, it proves necessary to distinguish between ‘progress in the human sciences’ and ‘progress in the natural sciences’. These issues, which pertain to the general philosophy of science, are treated in the final chapter of the book.

  16. That book sounds absolutely fascinating. I may need to find a copy. However, it seems to me that “the European tradition” can claim to have made one big advance over anything previous by the late 19th century: a theory of comparative historical linguistics. In general, comparative linguistics of any kind appears rather poorly developed in any period before the early modern era, and not just for lack of data or even lack of prestige of the languages involved: Aramaic, Hebrew, and Arabic would have been a great combination to build on. But maybe I’m underestimating early Indian work on that problem?

    Haven’t seen the video yet, but in general Al-Jallad does excellent and thought-provoking work on the history of the Arabic language.

  17. Al Jallad’s work on Ancient North Arabian is spectacular. What surprises me is that that work had not work long before. How would the European explorer-antiquarians of the 19th century miss kilometer upon kilometer of stones covered with inscriptions out in the open?

  18. David Marjanović says

    How would the European explorer-antiquarians of the 19th century miss kilometer upon kilometer of stones covered with inscriptions out in the open?

    Well, first you have to find those particular spots of not particularly interesting desert (“go straight ahead, then turn left in three weeks” as the joke goes), then you have to see the inscriptions (they don’t stand out that much in blinding light, I guess), and then you have to be able to read them (many of these scripts were not very widely used; of course they’re related enough that decipherment has been possible, but that seems to require knowing a good amount of them).

  19. Lameen: No, the Indian tradition (which the author argues, convincingly to my mind, was by quite a margin the most advanced pre-modern tradition of synchronic *descriptive* linguistics) was not especially advanced in the matter of comparative linguistics (If I recall correctly the closest it ever got to such a point involved some of the grammatical descriptions of various Prakrits, which often took Sanskrit as a yardstick for comparison, but without ever clearly understanding (or stating?) that Prakrits are later, changed forms of Sanskrit), so I think the author would agree with your assessment on the European tradition.

    We should bear in mind, mark you, that comparative-historical linguistics is a comparatively late, recent outgrowth of the European tradition. And one which, after all, arose as a direct result of European colonial expansion: the discovery by Europeans of Sanskrit was a major, if not THE major cause of the rise of comparative-historical linguistics.

    Hmm. I wonder. If history had taken a different turn and Western Europe had been colonized by South Asians/South-East Asians familiar with the Indian grammatical tradition, would these colonizers’ discovery of Latin and Greek (and/or some of the more conservative contemporary Indo-European languages of Europe) have triggered a similar birth of comparative-historical linguistics?

    I have to admit, the thought of comparative-historical linguistics using Pāṇinian descriptive techniques as a foundation is fairly mind-blowing…I am reminded of a science-fiction short story I read a long time ago, involving an alternate universe where a united Roman Empire was re-created a few centuries after the fall of the Western Roman Empire, and then went on to colonize much of the planet (including Eastern North America, where the story is set), with China its main global rival. Its technology is basically medieval, but it has a kind of computer technology whose foundation is “Pāṇinian” formal logic.

  20. > I have to admit, the thought of comparative-historical linguistics using Pāṇinian descriptive techniques as a foundation is fairly mind-blowing…I am reminded of a science-fiction short story I read a long time ago, involving an alternate universe where a united Roman Empire was re-created a few centuries after the fall of the Western Roman Empire. Its technology is basically medieval, but it has a kind of computer technology whose foundation is “Pāṇinian” formal logic.

    You might want to ask James Nicoll about that. He has an encyclopedic knowledge about such things.

  21. How would the European explorer-antiquarians of the 19th century miss kilometer upon kilometer of stones covered with inscriptions out in the open?

    They didn’t – they did notice the existence of Safaitic inscriptions, and got a fair way towards reading them. But they were looking at the region from a perspective centred on the great “civilisations” and the Bible, in which these short graffiti seemed very marginal. Al-Jallad brings a very different perspective to them, centred on the traditionally peripheral world of pre-Islamic Arabian nomads.

  22. Itkonen’s book sounds just like what I want.

    I recently completed a semester on the history of linguistics for which the set textbook was a horrible book (in Mongolian) by the writer and translator Ravdan. (Arabic linguistics was one part of the course, as was ancient Greece, Egypt, i.e., Alexandria, Rome, India, and China — Europe was not covered after Rome, so there was Donatus but no Priscian). The course was assessed on a ten-page written assignment for which I was assigned Port- Royal grammar.

    Naturally I started with Chomsky’s heavily criticised “Cartesian Linguistics”. Scholar after scholar debunked his claim that Port-Royal was Cartesian and that the Port-Royal grammarians were essentially proto-transformationalists and the first to propose a universal grammar. Robin Lakoff traced Port-Royal “deep structure” and “surface structure” to a textbook by Sanctius that taught how Latin authors achieved literary elegance through operations like ellipsis on underlying sentences that were simple but clumsy. (Chomsky dismissed Sanctius as being engaged in textual analysis, unlike Port-Royal, who were aiming for a psychological grammar.) Critics also pointed out that “universal grammar” as a concept could be traced back to the Modistae, with special reference to Roger Bacon, who claimed that languages only varied “accidentally”.

    Despite this trouncing, and despite having dropped both deep/surface structure and transformations in the meantime, Chomsky was still bringing up Port-Royal grammarians as the precursors of his modern Universal Grammar in 2013. (Apart from his stubborn unwillingness to admit he is wrong, unless it is part of his progress towards an ever purer expression of his computational vision of language, the only other explanation for Chomsky’s intransigence would seem to be that everything before Port-Royal was written in Latin, which, as Lakoff helpfully pointed out, Chomsky can’t read.)

    At any rate, this brought all kinds of pre-modern linguistics into view for me, and I would definitely like to know more about the Modistae, the Arabs, and the Indians. Except in the hands of arch-distorters like Chomsky, history is truly the key to the present. The $188 price tag on Itkonen’s book does give one pause, however….

  23. David Marjanović says

    I’ve now watched the video, and this one which sort of continues where that one left off, and this one which (unintentionally) sort of follows.

    In general, comparative linguistics of any kind appears rather poorly developed in any period before the early modern era, and not just for lack of data or even lack of prestige of the languages involved: Aramaic, Hebrew, and Arabic would have been a great combination to build on.

    And indeed their similarity was often remarked on – but an explanation was readily available: the Tower of Babel.

    Maybe reconstructing the last common ancestor of these languages would have been seen as reconstructing the pre-Babel language, which could have been construed as blasphemy.

    The similarity of Greek and Latin was remarked upon by a Roman or two, but interpreted as Latin having descended from Greek. I guess the other Italic languages were already extinct at that point…

    but without ever clearly understanding (or stating?) that Prakrits are later, changed forms of Sanskrit)

    …not that it matters, but they aren’t, not quite. Sanskrit (Vedic, Epic and Classical alike) had a megamerger of like six different PIE consonant clusters as kṣ; at least some Prakrits distinguished the originally voiced ones from the originally voiceless ones as ggh and kkh, so they can’t be descended from Sanskrit despite being descended from something very, very close to it.

  24. For my purpose it does not matter whether or not his analysis was historically or, for that matter, philologically, sound. Even if the ‘Cartesian linguistics’ as described in his book is just an artefact of his own views of the historical development of the discipline, it still stands as an interesting and relevant conception of the dichotomy between actual speech and underlying level, which may be used for comparative reasons.

    In the Arabic world a similar approach to language and thought was once current during a brief period …

    https://www.researchgate.net/publication/233659143_The_Notion_of_Underlying_Levels'_in_the_Arabic_Grammatical_Tradition

  25. an interesting and relevant conception of the dichotomy between actual speech and underlying level

    You don’t need to be a Chomskyan to understand that about language. Sanctius realised that before Chomsky but Chomsky denied the relevance of his insight. Do you agree with that too?

    Just as it doesn’t take much thought to realise that language and writing are independent of each other. You only need to be familiar with languages that have used several scripts during their existence (there are quite a few of them) to know that.

  26. In his conversations with Lakoff Chomsky denied Sanctius’ insight into underlying sentences and the “figured language” of elegant prose. If you agree with Cartesian Linguistics then you have to agree with Chomsky’s stance, too. Not a good position to take, even if it’s just “for your own purpose”.

    Perhaps you are speaking of Transformational Grammar, which set up a deep and surface structure, which is fine. But agreeing with the insights of TG should not mean agreeing with Cartesian Linguistics.

    Your reference to Greek elements in Arabic linguistic thinking was spot on, by the way. I found 29 pages available for view on Google Books and even from those 29 pages it was clear that the Arabs had some exposure to Greek grammatical studies.

  27. @Bathrobe: This reminds me of an old question I once had: Why were the Port-Royal Logic and Port-Royal Grammar written in French? The Abbey of Port-Royal-des-Champs was the center of Jansenism in the middle of the seventeenth century, which may naturally have attracted some heterodox thinkers. However, while Jansenism had a view of salvation very similar to Calvinism (emphasizing human depravity, irresistible grace, and predestination), the sect also preferred to maintain the Roman Catholic trappings of sacraments,* priesthood, and pomp—including the exclusive use of Latin for the liturgy. Certainly, the likes of Antoine Arnauld,** Blaise Pascal, and Pierre Nicole would have spoken impeccable Latin, but they chose to write the Port-Royal Logic and Port-Royal Grammar in the vernacular. It is probably relevant that Pascal’s several defenses of Jansenism were, I believe, also all written in French, rather than Latin.

    * I realize that don’t know what the Jansenist view on transubstantiation was, but I would guess that they believed in it.

    ** Antoine Arnauld was apparently the youngest of twenty children of his namesake father. Moreover, unlike Johann Sebastian Bach, who also famously had twenty children, the elder Arnauld had all his children with just a single wife.

  28. The elementary schools that Port Royal set up to teach children also taught in French, as opposed to the Jesuit practice of teaching in Latin. I’m not sure of the ideological background to this.

  29. unlike Johann Sebastian Bach, who also famously had twenty children, the elder Arnauld had all his children with just a single wife.

    That’s a low blow, I call ‘foul’! JSB’s first wife died unexpectedly, and whilst he was out of town (and the death was not connected to childbirth). He was distraught. “Erbarme dich” from the Matthew Passion is his very personal response.

  30. … and this [skype lecture] which (unintentionally) sort of follows.

    sort of 😉

    Good to know there are Qur’an nerds just as much as Torah nerds or Gospel nerds.

  31. Somehow my reading on Port Royal doesn’t seem to have prepared me for this discussion.

    But I’ll never forget its story of freedmen rebelling bitterly against being hired to grow and harvest cotton in tidal mud flats in which they sank up to their knees, a lesson I learned vividly when a camping-by-boat trip in the area with my cousin went awry.

  32. This Port-Royal/Port-Royal-des-Champs outside Paris (the one in France, not Texas).

  33. Bathrobe, sorry, it was a quotation from a paper by Versteegh (the link below).

    I was not making any point. I saw the title of this article before your posted your comment, the parallelism was rather striking.

  34. Sorry, drasvi, I saw but was a bit puzzled by the indentation and didn’t realise that Versteegh had written this.

    I’m disappointed to read Versteegh, writing in 1994, still referring to the Transformational Grammar that Chomsky was using in 1966, and generally adopting a dismissive attitude to criticism of the book (‘the reconstruction of Cartesian thought is not incontroversial’) by suggesting that criticism was mostly directed at Chomsky’s views being ‘coloured by his wish to demonstrate that there had been precursors for his own approach to linguistics’.

    In fact, the critics (including experts in that era as well as Lakoff, who was a transformationalist) level far more specific and substantive criticisms of Chomsky’s book. When people criticise Chomsky’s theories it appears virtually impossible for it to stick because of theoretical disagreements. But it’s disappointing that even when Chomsky spouts unsupported nonsense in an area completely outside his field of competence (as he does in his book), the same lies still seem to hang around years later, long after they have been completely debunked by well-informed critics.

    Versteegh does himself a disservice leading into the topic of underlying levels in the Arabic grammatical tradition with such comments.

  35. PlasticPaddy says

    @brett
    I have the impression that the Jansenists wrote in French in order to more easily reach a wider French audience and because they saw Latin as a symbol of an obscurantist and reactionary approach to theology, politics and education. In education they promoted the teaching of French and Greek, as well as Latin, which was an innovation that was copied generally.

  36. reconstructing the pre-Babel language, which could have been construed as blasphemy.

    depending on how early you like your “early modern”, it wouldn’t be blasphemy so much as having the answer without any need to reconstruct, i’d think. i mean, did anyone (within the traditions that embraced the ‘tower of babel’ myth of linguistic diversity as a curse) before dante or abulafia get any traction arguing against the idea that hebrew was the adamic tongue? (a genuine question – my impression is not, but it’s not my bailiwick)

  37. For a detailed working out of Sanctius’ so-called “transformations”, see the article Deep and Surface Structure Concepts in Renaissance and Medieval Syntactic Theory by W. Keith Percival (in History of Linguistic Thought and Contemporary Linguistics ed. Herman Parret. Walter de Gruyter, 1976 — the article is partly available online at Google Books).

    The parallels with Arabic practice are striking. Sanctius embraced ‘ellipsis’ as basic in understanding elegant Latin texts by good authors. Later linguists like to substitute ‘deletion’ for ‘ellipsis’ as it fits in better with Transformational theory, just as Versteegh translates idmâr as ‘deletion’. But idmâr could easily be translated as “ellipsis” without doing violence to the concept.

    The concept of “ellipsis” goes back to Priscian at least. It would therefore not be surprising if idmâr and “ellipsis” had similar roots, although I am not expert enough to assert this categorically. At most the similarities are suggestive.

    The Port-Royal Nouvelle méthode pour apprendre la langue latine (by Lancelot), a pedagogical grammar of Latin, was full of influence from Sanctius, including the usage ‘figurative sentence’ (which Chomsky calls ‘surface structure’). On the other hand, Chomsky’s beloved Grammaire générale et raisonnée adopted very little from Sanctius apart from some general formulations. This led Lakoff to conclude that the Grammaire générale had to be read in conjunction with the Nouvelle méthode in order to make sense. Which Chomsky pooh-poohed because Sanctius was nothing more than a “textual analyst”, unlike the Port-Royal grammarians, who were his true precursors. Lakoff’s 1969 review of the Grammaire générale et raisonnée, which appeared in Language and includes Chomsky’s dismissal of Sanctius, is also available online and is well worth reading.

  38. David Eddyshaw says

    It is an interesting review …

    It seems to me that neither what Sanctius was taliking about, nor what Versteegh describes as belonging to the Arab grammatical tradition, are really the same as Chomskyan levels at all. Unlike Chomsky, Sanctius and the Arab scholars were concerned with how meaning (assumed by Sanctius to be all to do with logic, and thus, in his Humanist view, common to all mankind) was expressed in the vagaries of any actual human language. This is a very different dichotomy both from Chomsky’s E-language versus I-language, or his quondam distinction between “deep” and “surface” structure. The entire Chomskyan project depends on pretending that meaning is some sort of epiphenomenon and can be largely ignored in syntactic analysis, rather than permeating syntax at all levels.

    All these things have in common is that they all involve “levels” of some sort – and that they’re all wrong.

  39. @Bathrobe, now when I am less sleepy, the full sequence:

    There is a book “Indigenous grammars across cultures” that I wanted but was not able to find. I do not know if it is good, but it has some chapters like “Uzbek indigenous grammar under the impact of Arabic lexicographers, Persian poets, and Russian schoolmasters” that sound intriguing.

    There is a thick book “History of the Language Sciences” (contents) whose first volume I began reading.

    There is a chapter 285 of this thick book by Versteegh (https://www.researchgate.net/publication/333036182_The_study_of_non-Western_linguistic_traditions). I only know some of the contents of the first book from this chapter (the book is not even in Google books…) and it is also where I learned about the second book. It also mentioned the book recommended by Etienne. Etienne’s recommendation reminded me about it, so I opened it.

    It has a section about pitfalls of translating medieval grammars (where he references the article quoted above).

    As I was exporing its bibliography (I wanted to find what else the author of the Uzbek acticle wrote) I also saw “The Notion of `Underlying Levels’ in the Arabic Grammatical Tradition”. And then I opened this window, and saw your comment.

  40. drasvi, I’m extremely grateful that you mentioned those books/articles. They are all tantalising / fascinating. It’s just that I’m not in the market for someone who says that Chomsky’s distortions don’t matter.

    In fact, as I mentioned, I suspect that both Port-Royal (via Sanctius) and the Arabs hark back to a single source: Byzantine/Latin/Greek grammar. Perhaps if Versteegh had looked a bit deeper he might have seen beyond Chomsky and found that possible connection. It certainly might have discouraged him from accepting Chomsky as providing “an interesting and relevant conception of the dichotomy between actual speech and underlying level, which may be used for comparative reasons”. A comparison between Chomsky’s distorted history and the Arabs would be far less interesting than finding an actual historical connection.

  41. Well, if I came up with a good way to disprove Chomsky and I knew that I would succeed and even become famous, I would rather dedicate my life to lechery.

  42. The man simply wrote too much.

  43. It’s not the quantity that’s the problem…

  44. Or to put it another way, if he had only written Syntactic Structures it would have been too much.

  45. Well, yes, not the quantity. I just do not know what’s the right word.

  46. I’m not talking about disproving Chomsky, just about Cartesian Linguistics.

    I don’t understand your aversion to disproving Chomsky. Would you consider yourself a staunch supporter? Or is it just due to a general desire to be contrary?

  47. J.W. Brewer says

    I don’t think valuing lechery more highly than Chomsky-debunking is contrarian. Surely it is the opposite ranking that is a rather eccentric niche interest.

  48. Ancient Indians probably should have been well equipped to invent historical linguistics, if they only had possessed the concept of histor(iograph)y. But the idea that the world has changed in major ways all across time, and that there is thus reason to document both the past and the present, is (1) not an obvious fact to human societies and (2) explicitly denied in the more conservative strains of Hinduism and Buddhism alike. An India that would have invented historical linguistics would have been in many ways, I think, unidentifiable as “India” as we know it. (I recall an essay from some years back that claimed that scientific historiography has only been developed twice: once in Greece, once in China.)

  49. I don’t understand your aversion to disproving Chomsky. Would you consider yourself a staunch supporter? Or is it just due to a general desire to be contrary?

    I agree with JWB: drasvi has given no cause to suppose he’s any kind of supporter of the Chompster, he simply has better ways to occupy himself.

  50. David Eddyshaw says

    I don’t think valuing lechery more highly than Chomsky-debunking is contrarian

    There is no genuine opposition here. A true Renaissance Man* does both.

    * Or Woman, naturally.

  51. David Marjanović says

    * I realize that don’t know what the Jansenist view on transubstantiation was, but I would guess that they believed in it.

    From the conspicuous lack of mention on Wikipedia I conclude the same. Anyway, the long French article mentions that a Jansenist translated the New Testament into French (and shows us the title page), so clearly the Jansenists wanted to do public outreach.

    Chomsky-debunking

    I haven’t read the paper yet, but this review and the Twitter thread linked from there strongly suggest it’s already been done – specifically, the poverty of the stimulus is disproven.

  52. David Eddyshaw says

    To be fair to Norbert Hornstein, accused at that link of being “Norbert Chomsky-is-never-wrong Hornstein”, what NH actually asserted was “Chomsky is never (stupidly) wrong.” I’m OK with that; ANC is a textbook example of how to be cleverly wrong. (It still amazes me how people who are evidently very clever indeed can still lack even basic commonsense. I’m evidently still young at heart.)

    Chomsky’s system is (of course) not actually susceptible to debunking. In its imago form, it is not vulnerable to falsification by counterevidence at all.

  53. Exactly. It’s like debunking Time Cube.

  54. Did anyone ever attempted proving this “poverty of stimulus” or at least formulating a falsifiable conjecture?
    If not, how it can be disproven?

  55. David Marjanović says

    “Grammars cannot be learned from realistic amounts of mere exposure to language; learners must have prior knowledge, or they’ll fail” may not have been intended as a falsifiable hypothesis, but as some sort of axiom.

    But it is a falsifiable hypothesis. Get an AI, which doesn’t and can’t have prior knowledge, to learn one grammar from such limited exposure, and the hypothesis is falsified.

    The AI in the paper learned a whole bunch.

    Note the contrast to the Time Cube, where it’s really hard to identify any hypotheses at all. (Or maybe I’m just EDUCATED STUPID.)

    In its imago form, it is not vulnerable to falsification by counterevidence at all.

    I don’t know if “all language is MERGE” is falsifiable, but it certainly looks that way…

  56. Did anyone ever attempted proving this “poverty of stimulus”

    I’m pretty sure M.A.K. Halliday disproved it (to the satisfaction of ordinarily intelligent people), in his studies of child language acquisition. But yeah the falsifiable conjecture bit is problematic: do you leave your baby naked on a rock to see if it acquires language? Compare it to a baby brought up by dumb mutes? Insist the parents use only subjunctive sentences of at least 12 words when tending the infant? All conjectures seem inhuman, even by Chomsky’s standards.

  57. ANC is a textbook example of how to be 𝑐𝑙𝑒𝑣𝑒𝑟𝑙𝑦 wrong.

    The quote, I’m sure, tells us more about Hornstein than Chomsky.

    (Focusing on him as a single scholar is a bit beside the point anyway; I don’t understand exactly what propelled him into such stratospheric fame (or even when exactly), but it’s clear it has gone beyond what one could possibly accomplish with just wits and charisma alone, and besides the man doesn’t really strike me as an academic Fabio either. But I’ll wait for the more informed hagiographies that are bound to be on the way in a few decades.)

  58. @J Pystynen: Are you showing off some kind of double italics in that quote there?

  59. David Marjanović says

    These are separate Unicode characters that are not in italics; the showing off pertains to knowing where to find them.

    Copypasta: 𝑐𝑙𝑒𝑣𝑒𝑟𝑙𝑦

    BTW, the last author of the new paper and of the Twitter thread is one of the two (!) last authors of the classic 2016 paper that analyzed a (small) corpus of Pirahã and found possible, but no clear, examples of recursion in it, unless of course every sentence of 3 words contains 2 operations of MERGE and thus 1 instance of recursion by definition.

  60. @DM, the principle should be formulated rigorously enough, so it would be possible to demonstrate that your AI is a valid example of a “learner” who “doesn’t have prior knowlege” and what it did is “learned grammar”. If it is vague, the likely outcome of an experiment is : “all handsome men wear lipstick” “my AI is handsome and it does not” “no, it is not what I mean by handsome”.

  61. a “learner” who “doesn’t have prior knowlege”

    wrt the ‘Hearing Faculty’ all humans have (or learn) an amazing ability to tell the direction a sound is coming from, merely from the timing difference in when it reaches one ear vs the other.

    Whatever neural/cognitive power that derives from: does that count as ‘prior knowledge’ in ANC’s sense?

    @drasvi I think your quest will never get beyond definitional difficulties. Even if it finds some operable definition, ANC is going to reject it as not what he meant — unless of course it can be demonstrated to support the ‘poverty of input’ hypothesis — in which case everybody else is going to reject the definition.

    It’s a rabbit-hole from which there’s no escape.

  62. David Eddyshaw says

    In its most sophisticated form (as I understand it), the Poverty of Stimulus argument is not really about there purportedly being too little positive input for the learning child to model its internal grammar on; the idea is that it is (logically?) impossible for the child to learn from the input that certain constructions are ungrammatical (a sort of variation on the theme of “you can’t prove a negative.”)

    This too seems not to stand up to actual investigation. But such inconvenient facts are of no importance. We already know the Answer.

    Unfalsifiability does not of itself mean that Chomskyism is inevitably valueless. I believe that the Great One Himself has revealed that the System is [now] what Popper would call a “metaphysical research programme” (“a possible framework for testable scientific hypotheses”) rather than a hypothesis; this expiicitly means that it is not liable to refutation in the way that a hypothesis is. [Popper was (rightly) unhappy with creationists perverting his labelling of Darwinism as being such a programme to imply that he thought that Darwinism was unscientific: that was pretty much the exact opposite of what he meant.]

    However, if Latter-Day Chomskyism is a metaphysical research programme (which seems fair enough) it is what Imre Lakatos called a degenerating research programme. In fact, pretty much a textbook example. Poster child …

    https://plato.stanford.edu/entries/lakatos/#ImprPoppScie

  63. @David Marjanović: That doesn’t quite fit with what I understand copypasta to mean. Granted, copypasta is a new term, and its meaning may not have completely settled. The first sentence of the Wikipedia article states:

    A copypasta is a block of text that is copied and pasted across the Internet by individuals through online forums and social networking websites [sic.].

    In particular, I think of copypasta as being a text that is sufficiently long that it would have to have been copied, rather than recreated de novo. Of course, it is permissible to make changes, but it should remain recognizable as a variant of the canonical version. (Just now, as I was searching for “gorilla warfare copypasta,” Google suggested that I might want to include the additional search term “clean.” Presumably that is for people who want a version without the initial, What the fuck did you just fucking say about me, you little bitch?) On the other hand, producing “𝑐𝑙𝑒𝑣𝑒𝑟𝑙𝑦” directly by using an alternative character set (rather than copying it, as I just did) might actually be deemed enough of a challenge to make “𝑐𝑙𝑒𝑣𝑒𝑟𝑙𝑦” a legitimate copypasta in the view of some.

  64. Sorry, arguments against POS have already been disproved by Janet Fodor: https://www.academia.edu/54881922/Understanding_stimulus_poverty_arguments?

    (And sorry, I could only read so far….)

    As for Norbert, I pretty much agree with the monicker he’s been given.

    The following exchange at Faculty of Language in 2012 sums up Norbert for me:

    A:
    Based on what you say here, you are in wide agreement with most possibly all players – certainly connectionists acknowledge built in biases of the way you describe. So can you be a bit more specific about what you think is ‘built in’

    Norbert:
    Read reply to Alex. There I built in most of principle A-C of binding theory, c-command and locality domains. Them’s my biases.

    A:
    Thanks. Just so I understand most of principle A-C of binding theory, c-command and locality domains are encoded by the human genome and instantiated in our brains? I am not asking how and where [as i assume we do not know that yet]only that. And they are in the brain of every human not just speakers of English, right?

    Norbert:
    Yup. I build in the binding theory, or the functional equivalent thereof (I actually think that most of BT can be reduced to the theory of movement properly construed, but this does not change the nature of what is built in substantially). I don’t know if it is in the genome, but it is NOT learned. It’s what learning presupposes. They are in the brain of everyone: part of FL. It is another example of the POS, but this time with more cross linguistic support than Y/N questions which is not nearly as widespread a phenomenon.

    A:
    Thanks. You say: I don’t know if it is in the genome, but it is NOT learned. – Does this mean there are further options? If so what are they?

    Norbert:
    Damn if I know. Epigenesis, God, Paul Krugman when he isn’t battling the Republicans? Does anyone really know how things like learning biases get coded in the genome? Does anyone know how the bee dance or the dead reckoning behavior of ants gets coded in the genome? But, seriously, I don’t know, or, right now, care. That question is way above my pay grade.

    @ drasvi:

    I’m not that interested in disproving Chomsky. It would take a lifetime and require far more smarts than I have. My point was that people keep on believing Cartesian Linguistics long after it’s been disproved. I couldn’t understand why you would come out with such a comment (“if I came up with a good way to disprove Chomsky and I knew that I would succeed and even become famous, I would rather dedicate my life to lechery”) out of the blue, when it wasn’t related to what I was saying.

    At any rate, the papers you linked to or mentioned are gold. Lakoff, Percival, and now Versteegh suggest a tentative but fascinating linking of the dots. Not that I have the necessary knowledge of Latin, Greek, or Arabic to actually go in and prove the links….

  65. drasvi: the principle should be formulated rigorously enough, so it would be possible to demonstrate that your AI is a valid example of a “learner” who “doesn’t have prior knowlege” and what it did is “learned grammar”.

    AntC: I think your quest will never get beyond definitional difficulties.

    Since Chomskyanism maintains that the human brain is a computer the operations of which can be described mathematically [Chomsky boils it all down to Merge], I should think that Chomskyans would have a hard time denying the relevance of AI. (I personally wonder whether the operations of the human brain can be modelled mathematically, but I have encountered people who believe exactly that. And I am, again, not in a position to disprove them.)

  66. @DE it is (logically?) impossible for the child to learn from the input that certain constructions are ungrammatical

    You mean that for some posited construction, the learner can’t distinguish whether its from the infinite set of grammatical but not-yet-observed sentences or the infinite set of ungrammatical sentences(?) (They’re quite likely to observe ungrammatical sentences, because people often make slips of the tongue/false starts/etc.)

    But kids _do_ learn (try out) constructions that are ungrammatical — typically they over-generalise; then they learn ‘this verb is irregular’, etc. How? Typically an adult or peer will correct them.

    They could observe an irregular verb form from another speaker in a context where they would posit the regular. They could observe a speaker corrects themself after a slip of the tongue.

    So does this POS argument mean a learner can (logically possibly) build only from examples, not from meta-data like corrections or awkward reactions amongst hearers and speakers?

    How stupid does ANC have to hypothesise learners are, that they willfully impoverish the Stimulus?

  67. David Eddyshaw says

    He’s cleverly stupid. (There’s a lot of it about.)

    I perhaps should make clear that I don’t subscribe to any of this myself. I’m just trying to identify the Chomskyans’ better arguments*, though I find it difficult to muster up enough enthusiasm to do so very effectively, as the entire enterprise strikes me as essentially vitiated from the get-go by fundamentally mistaken ideas about the nature of language, which mean that any actual worthwhile results (there have been some, on a modest scale) occur despite the Theory, not because of it.

    * I am quite taken aesthetically by beautiful reductionist systems which I don’t believe in at all, like Lucretius’ Epicureanism (especially) or Wittgenstein’s Tractatus (even Language, Truth and Logic has its appeal on this level.) Unfortunately, Chomskyan UG manages to be a remarkably ugly reductionist system, combining a singularly content-free and boring core with a horribly messy and evidently largely arbitrary superstructure unconvincingly asserted (retrospectively) to be a rigorous consequence of it.

  68. I am quite taken aesthetically by beautiful reductionist systems which I don’t believe in at all, like Lucretius’ Epicureanism (especially) or Wittgenstein’s Tractatus

    Ah, I see we are of the same cast of mind. The Tractatus is indeed as beautiful as it is wrong. Philosophical Investigations has a beauty that must be wrung out of it by taking a long hike to digest each observation.

    Then I have to confess I found Syntactic Structures beautiful in the way of the Tractatus, and everything ANC has written since as progressively debasing the coin.

  69. It seems to me that neither what Sanctius was taliking about, nor what Versteegh describes as belonging to the Arab grammatical tradition, are really the same as Chomskyan levels at all. Unlike Chomsky, Sanctius and the Arab scholars were concerned with how meaning (assumed by Sanctius to be all to do with logic, and thus, in his Humanist view, common to all mankind) was expressed in the vagaries of any actual human language.

    @DE from the same paper: In the commentaries the main effect of restitution of the deleted element is the reconstruction of the intention of the speaker. In the Kitâb, on the other hand, deletion is an essential factor in the analysis of the formal, syntactic relations between elements of the sentence, and the reconstruction of the deleted element is necessary in order to explain the effects of governance (ʿamal) in the sentence.

  70. @Bathrobe, I meant that I would have to dedicate a good chunk of my life to Chomsky.

    I would not expect criticism of Chomsky from a philologist (an Arabist in this case). I do not see how he can be attractive for them, a serious criticism of even one book takes time (better spent on a more appealing subject like: medieval cookbooks) and he is a parallel universe (next door on the same floor in the university). I can see why one would limit himself to a constatation that a certain view exists. As to why mention Chomsky at all… we inevitably read ancient and foreign grammars in the context of modern/local theories. It is always tempting to say “this Klingon theory is like our tagmemics!”. It’s understandable…

  71. David Eddyshaw says

    @drasvi:

    Thanks. Look like I was wrong, at least about the clever Arabs.

  72. @DE, actually, anyone who needs “virtual” elements in Arabic grammar can start with the alif.

  73. Since Chomskyanism maintains that the human brain is a computer the operations of which can be described mathematically [Chomsky boils it all down to Merge], I should think that Chomskyans would have a hard time denying the relevance of AI.

    Yes. But then the theory must look this way:

    (1) a formal model
    (2) a theorem about this model.
    (3) a claim that real human langauge learning can be approximated by the model.

    Then someone would demonstrate that the theorem is false. But I suspect they do not have a model.

  74. David Eddyshaw says

    @drasvi:

    Surely alif is just the symbol for /ʔ/, which is a perfectly good consonant in Classical Arabic? Admittedly alif is also used to mark vowel length, but then so are the symbols for /j/ and /w/.

    The other complications in spelling just arise from the fact that Muhammad’s own dialect mostly dropped /ʔ/, so it had to be reinserted into the text of the Qur’an using diacritics.

  75. I will better give a link: p 221. (I am NOT promoting Kees Versteegh)

  76. David Eddyshaw says

    It looks like the grammarians in question were misled by the (Aramaic-derived) orthography of Arabic long vowels* into supposing that they were in fact VC sequences – including /a:/. In fact they seem to have denied the existence of long vowels altogether, which inevitably led them to posit imaginary consonants. This can be made to work … and to be fair, there really are similarities in the behaviour of CVC and CV: syllables in Classical Arabic, so it’s not a daft idea.

    The price is that you end up with an imaginary consonant which (unlike /w/ and /j/) appears exclusively in the two contexts: a_C and a_#. It also doesn’t sit very well with Arabic morphophonemics, as you end up having to insert extra consonants into the triliteral consonantal templates all the time (though that mightn’t be such a bad thing given the way “hollow verbs” and the like actually behave.)

    * And presumably by the loss of word-internal /ʔ/ in the dialect underlying the consonantal orthography of the Qur’an.

  77. David Eddyshaw says

    There are tantalising hints that Oti-Volta long vowels all either go back to original VC sequences or are due to secondary lengthening of short vowels in stressed open syllables, but if so, there has been so much levelling and remodelling of paradigms that it’s difficult to be sure.

    And PIE has its famous laryngeals …

  78. @DE, I am not sure about “misled”.

    Whether you read -iiC and -ijC the same or differently depends on your phonology. Practically I can’t make sense of derivation (rather than conjugation) without identifying an element of a long vowel with a glide.
    Also what you said about “similar behaviour” (not necessarily CA).

    But I am speaking about y and w here.

  79. From Al-Jallad’s book chapter that D.E. linked to (Section 3.1)

    “The evidence suggests that the prefixed article *han- emerged in the central
    Levant sometime in the late second millennium BCE, after the diversification of
    Northwest Semitic …”
    “All of these languages, including the earliest Old Arabic, took over the form
    of the article unchanged; that is h- with the assimilation of the /n/ before a consonant,”
    “… the earliest linguistic stages of Arabic – and indeed
    Proto-Arabic – lacked a fully grammaticalized definite article. Contact with
    Canaanite then seems to be the likeliest explanation for the appearance of the
    h-article in Old Arabic.”
    “While the h- article is the commonest form in Old Arabic, whence the ʔal form?
    The ʔal article appears to be a later development from the original han article,
    through two irregular sound changes: h > ʔ and n > l.”

  80. @ J Pystyren : Richard Carrier

  81. David Marjanović says

    Well, most long vowels in German go back to /Vx/ sequences, lengthening of short vowels in stressed open syllables, and/or lengthening of short vowels in monosyllabic words… and similarly in English…

    The ʔal article appears to be a later development from the original han article,
    through two irregular sound changes: h > ʔ and n > l.

    Not impossible, but articles do get replaced sometimes. A large chunk of Romance seems to have switched from a descendant of ipse to one of ille.

    The first sentence of the Wikipedia article states:

    A copypasta is

    Huh. I had never even seen a countable copypasta.

    Popper was (rightly) unhappy with creationists perverting his labelling of Darwinism as being such a programme to imply that he thought that Darwinism was unscientific: that was pretty much the exact opposite of what he meant.

    If “Darwinism” means “evolution by mutation and selection”, it’s pretty easily falsifiable.

    (…and indeed has been – you have to add “and drift”.)

    For example, creationist frontloading (as accidentally revealed in that infamous Star Trek episode) predicts genomes that could not withstand mutation, selection or drift. If they existed anyway, that would falsify a whole lot.

  82. The ʔal article appears to be a later development from the original han article,
    through two irregular sound changes: h > ʔ and n > l.

    “Two irregular sound changes”: that can happen, but takes good evidence. How are these sound changes supported?

  83. David Eddyshaw says

    articles do get replaced sometimes

    All the Western Oti-Volta languages have definite articles; but although WOV is by and large a bit less internally diverse than Romance, these definite articles are of several different origins, sometimes even within subbranches (Kusaal differs from Mampruli, for example); most derive from postposed *la “that” *ŋ͡ma “this” or *maga “only”, but the Dagaare dialect chain has gone its own way with a proclitic a, probably from *ɲa, if it’s related to the particle that precedes personal names elsewhere in WOV, which seems to be the only real parallel.

    The oddity is not so much that there are so many different-origin definite articles, which is easy enough to explain on the hypothesis that Proto-WOV, like Latin, didn’t have one at all, but that all the languages do have a definite article. It tempts you to speculate about substrates … (on the other hand, lots of planets have a north. I mean, lots of languages have independently created definite articles from other elements, so maybe casting around for “explanations” is a bit pointless anyhow. Or maybe it was just a case of definite-article envy between WOV-language speakers. “Hey, that‘s a cool way of organising discourse and producing textual cohesion. Let’s do that, guys!”)

  84. David Eddyshaw says

    I wonder if there are any languages that have lost definite articles altogether in the course of time? If there aren’t (or if the exceptions can be handwaved away) it might be another of those alleged diachronic ratchets like SOV -> SVO word order that supposedly only ever go in one direction. And we can PROVE that Proto-Human had no definite article! Yes, PROVE, I say!

    [Actually, I think there are; the definite article gets denatured with time and, although still present, no longer marks definiteness as such. It’s happened with Aramaic and with Mande, now I think of it. And even Niger-Congo noun-class affixes may have started out as articles once upon a time. (Best Beloved. Do you see?)]

  85. “Two irregular sound changes”: that can happen, but takes good evidence. How are these sound changes supported?

    @Y you’ll have to read Al-Jallad for yourself (I am not worthy to wash his feet). I elided a long explanation; and there’s a longer explanation forthcoming in a not-yet-published (as of 2019) paper. Could be this from 2021, or a precursor.

    “The former is well attested
    in Arabic (e.g. the causative ʔaCCaCa from haCCaCa), while the latter is
    not uncommon in loans (e.g. finǧān vs. finǧāl ‘cup’). The ʔal article appears to
    have developed in the western dialects of Old Arabic, attested first in the Nile
    Delta (cf. the famous αλιλατ al-ʔilat ‘the goddess’ mentioned in Herodotus, Histories
    I: 131), and is the regular form of the article in the dialect of the Nabataeans, …”

    Herodotus ‘innit.

  86. If “Darwinism” means “evolution by mutation and selection”, it’s pretty easily falsifiable.

    What boggles my mind still with Darwinism is the Poverty of the Stimulus at the time he published: a very sketchy fossil record, with huge gaps in it; no carbon dating; no evidence/mechanism for continental drift; no gene theory; no knowledge of DNA. He got there from birds’ beaks in the Galapagos.

    Has Chomskyanism delivered any productive/falsifiable hypotheses? The whole project seems to be devoted to withdrawing from making any falsifiable claim into more and more ineffable abstractions.

  87. I wonder if there are any languages that have lost definite articles altogether in the course of time?

    Possibly, mine.
    https://en.wikipedia.org/wiki/Proto-Slavic_language#Adjectives

  88. @David Marjanović: There are certainly situations in which it is useful to think of a gene pool changing through a drift process. However, actually trying to distinguish rigidly between genetic drift and selection results in a mess of unfalsifiable nonsense.

  89. David Eddyshaw says

    Possibly, mine.

    Another good example (again, definite article not so much dropped as bleached.)

    At this rate we will soon be proving that Proto-Human DEFINITELY had a definite article …

    I had a vague memory that Ki-Nubi (which is an Arabic creole) had no articles, but I just looked it up and I was wrong. It’s replaced the Arabic definite article with a postposed de. (Much like Haitian, which has a postposed la in place of the French preposed articles.) Creoles seem to like articles (though not Lingala, or Bislama …)

  90. David Eddyshaw says

    What boggles my mind still with Darwinism is the Poverty of the Stimulus at the time he published

    Indeed. We are remorselessly driven to the conclusion that Darwin possessed an innate Theory of Evolution organ. It probably arose as a result of a single mutation. This is consistent with the genetic evidence:

    https://en.wikipedia.org/wiki/Darwin%E2%80%93Wedgwood_family

  91. @drasvi
    I meant that I would have to dedicate a good chunk of my life to Chomsky.

    Then we are on the same page.

    My only aim in reaching a deeper understanding of Chomsky is to excise him from my life. DE seems to have reached a healthy position. Chomskyanism seems to believe that it is approaching the deepest secret of the ability of humans to do language while leaving out virtually everything that makes language interesting (differences iin vocabulary, usage, sociology, etc.). I am not willing to devote my life to his theories. A life of lechery would indeed be preferable.

    I was, however, referring only to Cartesian Linguistics, where Chomsky got it laughably wrong, although his abstract style (a kind of beating around the bush to assert that his own opinions are “on the right track”) makes it sound like he has something to say.

    I would still recommend reading the Percival paper in conjunction with the Versteegh paper. It seems to me there is a lot in common.

  92. Darwin possessed an innate Theory of Evolution organ. … arose as a result of a single mutation.

    That gene must have been dominant over the Exquisite Pottery organ — because even marrying a Wedgwood cousin didn’t produce any potters.

    I see a Keynes in the bottom right of that family tree. Wherefrom the General Theory of Employment Interest and Money organ?

    I’m beginning to doubt your “cleverly stupid” explanations. “Stupidly stupid” seems a more parsimonious hypothesis.

  93. David Marjanović says

    Or maybe it was just a case of definite-article envy between WOV-language speakers. “Hey, that‘s a cool way of organising discourse and producing textual cohesion. Let’s do that, guys!”

    Well, the Germanic definite articles have been blamed on Romance, the Romance ones on Greek, and I think even the Greek ones on Egyptian… also, everything in contact with the Roman Empire (Brythonic, Albanian, Berber) seems to have definite articles, even if that’s where the similarities end (e.g. Basque has postponed ones despite being surrounded by Romance languages with preposed ones).

    And there was a previous round. The “long forms” of Slavic adjectives are shared with Baltic, where they’re apparently still definite, and Germanic next door also uses special adjective endings to go with definite articles, unless endings are generally lost. The Germanic ones are from *n-stem nouns, the Balto-Slavic ones from the PIE relative pronoun that Germanic just lost without a trace I can think of.

    PIE marking the animate nominative singular with *-s rather than zero is a typologically odd thing, and suspiciously similar to the animate nominative singular demonstrative pronoun *so; maybe it was a postposed definite article at some point. There’s even independent evidence for a round of apocope in pre-PIE.

    He got there from birds’ beaks in the Galapagos.

    and the rest of this world tour and pigeon breeding and chicken breeding and reports of swimming bears and an enormous amount of barnacles. 🙂 The Origin is a large book, and much of it consists of presentations of data.

    However, actually trying to distinguish rigidly between genetic drift and selection results in a mess of unfalsifiable nonsense.

    Really? A driven trend looks quite different from a fake trend that is caused by starting next to a wall. Or maybe I’m thinking in much longer timeframes than you here?

  94. Postponed articles
    And then there is the Balkan Sprachbund, where most members (Bulgarian, Macedonian, Romanian, Albanian) have developed postponed articles, despite what is usually seen as the original instigator (Greek) having a preponed article.

  95. Ki-Nubi … replaced the Arabic definite article with a postposed de. (Much like Haitian, which has a postposed la in place of the French preposed articles.)

    bzzzt There appear to be two syntaxes so your the in ‘Arabic definite article’ should not be a definite.

    From Al-Jallad’s 2021 paper linked above, section 2 ‘The History …’

    “Overt marking of definiteness is not reconstructable for any major node of the Semitic family. The earliest attested Semitic languages lack this feature – it is absent in Ugaritic, Amarna Canaanite, Gəʕəz, and is rare in the earliest stages of Hebrew and Aramaic as well.
    The article, therefore, seems to have emerged sometime in the middle of the 2nd millennium BCE and diffused areally among the Central Semitic languages.
    “All of the Central Semitic definite articles seem to derive from a single demonstrative/presentative element, *han, which grammaticalized along two patterns: in Canaanite and Arabic, the article is a prefix, while in Ancient South Arabian, Aramaic, and Hasaitic, it is a suffix.” [TROPPER (2001), pace RUBIN (2005). Seems to be in Journal of Semitic Studies Vol XLVI, in German]

    So did Ki-Nubi get its postposed definite from South > Sudanese Arabian?

    Did (South) Arabian influence get to Volta-Congo before the proto-Haitians left those shores? We’d have to be careful about the timing of any contact [Al-Jallad again]

    “For reasons that we can only speculate about, the al-
    article ended up as the dominant form by the 7th c. CE.”

    I can’t see any explanation for when/if the suffix form died out — Al-Jallad’s chiefly accounting for the presence of the nasal in South Arabian. (I am by now in the weeds well over my pay scale.)

    Oh, and @Y yes the History is horribly diffuse; there’s plenty of examples of h- co-existing with ‘- and -n/-m co-existing with -l.

  96. @David Marjanović: As I said, there are plenty of situations where drift is a reasonable model. However, to try to demarcate an actual line between changes in the gene pool that are driven versus undriven requires one to postulate some sort of idealized fitness function that may differ from what actually determines whether actual individual organisms survive. You can try to argue that some survived to reproduce because they were particularly fit, while others survived purely via chance. However, attributing things to chance is really no more than an admission that the particulars of why a specific organism did or did not prosper are unknown to us; however, with finer-grained knowledge, we could potentially answer the question of exactly why some performed better than others.

    For example, it may be that between two otherwise nearly phenotypically identical animal siblings, one was a bit slower than the other. Under a broad range of circumstances, this would tend to me the slower sibling less likely to survive, but if the faster one’s speed happened to cause them to blunder into the reach of a predator, the slower has turned out to be more fit. This was not a predictable form of fitness, since it depended on a one-time contingency of the animals’ precise environment. Yet it is a form of fitness nonetheless, different only quantitatively in scale and predictability—not anything qualitative—from more obvious manifestations of individual fitness.

    Put another way, “survival of the fittest”* is tautological, which is something that a lot of evolutionary biologists do not like to hear, (ostensibly**) because it makes it sounds like the whole edifice of Darwinian evolution is tautological. However, the full evolutionary paradigm is not tautological, because of what we understand about how organismal variability is inherited. Moreover, there are lots of situations when it is possible to understand what specific factors tended to make organisms more or less likely to survive and to pass along their genes; it is perfectly reasonable to study “big picture”*** fitness functions that operate over long time scales and sizeable populations.

    * In 1992, a writer from Oregon named Robin Cody wrote a book called Ricochet River. It was, in many ways, a remarkable encapsulation of the Pacific Northwest zeitgeist of the time, and some critics seriously compared it to Catcher in the Rye. The story is set in 1960, but it was about all the things that seemed particularly important in Oregon during the last two decades of the twentieth century: the decline of the logging industry, the changing roles of American Indians, and the problems plaguing the salmon. I attended an appearance by Cody, in which he first talked about how much research he had put into understanding the environmental issues surrounding the salmon, before reading a couple scenes from the book, including one that involved one of his main characters trying to throw salmon whose progress upstream had been blocked by a dam up and over the barrier, so they could reach their spawning grounds.

    (Salmon have a unusual life cycle, which made them particularly challenging to conserve. They are born in freshwater streams, then travel downriver to the ocean, where they live for several years, before returning to their freshwater origins to spawn and then die. The vast numbers of dead salmon provide a big nutrient influx into the stream ecosystems, much of this eventually finds its way through the food web to their growing progeny. With the damming of many of the rivers in the Columbia River watershed for electricity, it became difficult for the salmon to return upstream to spawn. The dams were built with fish ladders, which enabled the salmon that located them to bypass the dams and reach their spawning grounds, but that did not really allow enough fish to make it back upriver. The key reason for this failure was not initially appreciated, since it is related to another unusual feature of the salmon lifecycle. Each salmon tries to return, based on scent memories, to the very stream where it was hatched, which means that every little brook was home to an essentially separate genetic population. There was some crossing over of genetic material, since some fish would take wrong turns, but for the most part, each headwater was its own isolated gene pool. That meant that the number of fish that needed to get upstream every year to keep the populations healthy was much larger than was originally estimated, since there needed to be a sufficient breeding population in practically every creek. Many salmon runs had gone extinct by the 1990s, either because too few were making it back to their home waters, or because those home waters were rendered unfit for spawning by other forms of human activity. Given the economic and ecological importance of the salmon species, it was a real challenge to rebuild the populations and figure out ways to make the populations viable in the long term. This led to a lot of contentious politics in the Pacific Northwest in the 1980s and 1990s.)

    At that reading by Robin Cody, the author talked about how he wanted to portray how the construction of the dams had led to an inversion of “survival of the fittest,” because the weakest and most timid fish—the ones who had never come down from their headwaters to the sea—would be the only ones able to breed. I was stunned that, despite all his apparent research, Cody appeared to think that there were salmon who could live and thrive without ever traveling down to the ocean, living instead in freshwater all their lives. It was a remarkable display of ignorance on a topic that he had supposedly devoted so much time to trying to understand—although I later came to wonder how many other people who had opinions about what should be done to address the salmon problems were equally poorly informed about the fish’s biology.

    ** Many years ago, I used to read and comment at Larry Moran’s blog. Moran is an evangelist of genetic drift as a major force in evolution. In one comment, I laid out my argument that a sharp division between selection and drift is not really possible, and he got furiously angry. He called me a monster, for stating that those organisms that survived and reproduced were, ipso facto, the most fit. That implied, after all, that the Jews murdered by the Nazis were “less fit”; and did I really mean to argue that? I pointed out that he was applying the fallacy of equivocation, and, moreover, that I had no problem confirming that, since Darwinian “fitness” was not a quality with any moral valance, my relatives who had died in the Shoah had—in that particular sense of the word—been less “fit.” However, I was so disgusted with Moran’s behavior after this exchange that I have had nothing to do with his site every since.

    *** However, my senior quote in Technique (the MIT yearbook) was, “Don’t get too caught up in the big picture.” I was surprised by how many people read that and failed to notice that it was an inverse of the usual maxim.

  97. David Eddyshaw says

    So did Ki-Nubi get its postposed definite from South > Sudanese Arabian?

    No. It’s a creole of Regular Arabic. Sudanese spoken Arabic is not “South Arabian.”

    By South Arabian, Al-Jallad does not (despite the confusing name) actually mean any form of Arabic. It’s a separate branch of Semitic (like Modern South Arabian, which seems to be yet another separate branch.) It’s called “South Arabian” entirely because of where it was spoken; Actual Arabic is, if anything, closer to Canaanite/Aramaic/Ugaritic etc (which would make sense if it actually comes from the adjacent area, as Al-Jallad is implying.)

    The kind of Arabic which has undoubtedly affected the vocabulary of WOV languages (largely, if not entirely, indirectly, via other languages) is similarly Arabic-Arabic, whether Classical or one of the spoken forms.

    As far as I can tell, there is no evidence for any syntactic influence from Arabic on Oti-Volta at all. The languages work very differently: for example, Oti-Volta possessors invariably precede the possessum, there are postpositions (though also typically one or two prepositions as well to keep the typologists guessing), usually at least six genders instead of a measly two, very productive compound noun formation … the Oti-Volta protolanguage may even have been SOV, depending on what you make of the fact that personal-pronoun objects always precede the verb in Eastern Oti-Volta. It would go with the fact that the languages in other respects tend to go typologically with SOV languages rather than SVO (for what little that is worth), apart from their characteristic-yet-bizarre compounding of head nouns with all adjectives and even with demonstratives, which I gather the Chomskyites have proved by unassailable logic to be completely impossible.

  98. including one that involved one of his main characters trying to throw salmon whose progress upstream had been blocked by a dam up and over the barrier, so they could reach their spawning grounds.

    That reminds me of Yuri Nagibin’s wonderful story На тихом озере [On a quiet lake], in which the protagonist similarly tries to rescue eels in a Belorussian eel-fishing village.

  99. No. …

    Uh, ok thank you. I’ll stop trying to throw red herrings upstream.

    And since our host has just put in an appearance: thank you Hat for stimulating such a wide-ranging thread. Al-Jallad is “lively and informative” in writing as well as in video.

  100. Stu Clayton says

    @Bademantel: … where Chomsky got it laughably wrong, although his abstract style (a kind of beating around the bush to assert that his own opinions are “on the right track”) makes it sound like he has something to say.

    That’s an illuminating characterization. It is applicable as well to texts I am occasionally confronted with at work. “Pompous, tumescent Managerdeutsch” was the only description I had come up with so far. You have provided me with new analytical tools.

  101. So did Ki-Nubi get its postposed definite from South > Sudanese Arabian?

    The form de is from the usual Egyptian and Sudanese Arabic demonstrative da “this”, and is postnominal as in both.

    characteristic-yet-bizarre compounding of head nouns with all adjectives and even with demonstratives

    I don’t generally transcribe Korandje that way, but I don’t see anything in the speech signal to rule out such an analysis. One doesn’t pause between the noun and the adjective, as far as I’ve noticed.

  102. Stu Clayton says

    @Brett: However, attributing things to chance is really no more than an admission that the particulars of why a specific organism did or did not prosper are unknown to us; however, with finer-grained knowledge, we could potentially answer the question of exactly why some performed better than others.

    @Laplace:
    # We may regard the present state of the universe as the effect of its past and the cause of its future. An intellect which at a certain moment would know all forces that set nature in motion, and all positions of all items of which nature is composed, if this intellect were also vast enough to submit these data to analysis, it would embrace in a single formula the movements of the greatest bodies of the universe and those of the tiniest atom; for such an intellect nothing would be uncertain and the future just like the past would be present before its eyes.
    — A Philosophical Essay on Probabilities[3] #

  103. Arab: relating to the people or the culture.
    Arabic: relating to the language or the culture.
    Arabian: relating to the geographical region, i.e. the peninsula.

  104. there were salmon who could live and thrive without ever traveling down to the ocean

    Massive rapids at the head of the Kronotskaya River prevent fish from entering or leaving the lake. Those that live in the lake are therefore of special scientific interest as model studies in microevolution processes. They comprise a population of landlocked sockeye or kokanee salmon and a group of char distinguished by its significant polymorphism and plasticity: researchers have identified between three and five different forms.

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

    Genetic Differentiation of the Resident and Anadromous Sockeye Salmon Populations of the Kamchatka Peninsula: An Evolutionary Scenario for the Origin of the Resident Sockeye Salmon in Lake Kronotskoye

    https://www.researchgate.net/publication/339161133_Genetic_Differentiation_of_the_Resident_and_Anadromous_Sockeye_Salmon_Populations_of_the_Kamchatka_Peninsula_An_Evolutionary_Scenario_for_the_Origin_of_the_Resident_Sockeye_Salmon_in_Lake_Kronotskoye

  105. Always to be stressed as Árabic, not Arábic (I hear the latter pronunciation remarkably often in Europe from L2 speakers ).

  106. thank you Hat for stimulating such a wide-ranging thread.

    Don’t thank me, thank the voluble and informative Hatters! I never have any idea which post is going to inspire a thread like this; I expected maybe half a dozen comments. I just run ’em up the flagpole and see who whips out their lighters.

  107. David Eddyshaw says

    I don’t generally transcribe Korandje that way

    In Oti-Volta it’s very clearcut, because the first element of a compound loses its class suffix whether it’s a head or a dependent, e.g. Kusaal

    bʋʋg “goat”, bʋʋs “goats”
    kugir “stone”, kuga “stones”

    but

    bʋpiel “white goat”, bʋpiela “white goats” (or bʋpiellig, bʋpielis)
    bʋkan “that goat”, bʋban “those goats”

    kugpiel “white stone”, kugpiela “white stones” (or kugpielig, kugpielis)
    kugkan “that stone”, kugban “those stones”

    just like

    bʋkuos “goat-seller”
    kugkuos “stone-seller”

    (The variant adjective forms are a relic of noun-class-based grammatical gender agreement, but they are now basically in free variation.)
    Standard Kusaal orthography actually writes all these as single words, except when the noun combining form happens to look like a singular segmentally:

    dau “man”, dap “men”
    dau kan “that man”

    This is basically a mistake, though; the forms are in reality distinct: dāu “man”, dàu kàn “that man”, and dàu “man-“, just like “goat-” and kug “stone-“, cannot stand in isolation but is bound to the right.

    (To complicate things, many Kusaal nouns have actually remodelled the combining form segmentally on the singular [as, in fact, has happened with dau], but interestingly this never extends to remodelling the tone, which remains that of the expected suffixless form. Combining forms are also followed by different tone sandhi from singulars, whether remodelled or not.)

  108. “By South Arabian, Al-Jallad does not (despite the confusing name) actually mean any form of Arabic. It’s a separate branch of Semitic (like Modern South Arabian, which seems to be yet another separate branch.)”

    To complicate the terminology more, in AntC’s quote it is Ancient South Arabian (which is NOT an ancient form of Modern South Arabian).

    (Wikipedia complicates it even further: “The Ancient South Arabian script branched from the Proto-Sinaitic script in about the 9th century BCE. It was used for writing the Old South Arabian languages …“)

  109. And there is MSA as used by language educators and their students and MSA as used by linguists.

    I do not like the former, I am not against the latter, but it does not make it any less confusing.


    “Árabic”
    Yes, in Russian it is aráb, arábskij.
    So in my head I often say “Arábic” and I do not know how I usually stress it in reality.

  110. David Eddyshaw says

    @drasvi:

    “Old South Arabian” seems to be the standard name in English:

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

    It (or “they”, there being more than one language) is closer to Arabic than to the Modern South Arabian languages, in that there’s enough evidence there to show that it had the innovating yaqtulu imperfective, like Arabic along with Canaanite/Aramaic/Ugaritic, and unlike MSA and Ethiosemitic.

    In any case, as you say, it’s not ancestral to Modern South Arabian. These guys need to invent some new terminology …

    [I just noticed that the WP article compounded the confusion further down by wrongly calling it “Old South Arabic” twice. I was sufficiently motivated today to fix it.]

  111. And then there is a discussion of whether people of Yemen (during the rise of Islam or now) are “Arabs” or not:-E

    “…depending on what you make of the fact that personal-pronoun objects always precede the verb in Eastern Oti-Volta. …”

    As in Europe…

  112. An adjective “Arabian” was used much more widely in the days of Hagarians and Muhammedans (before Muslims, after Saracens).

  113. David Eddyshaw says

    @drasvi:

    In re Arabic long vowels:

    I agree that there is no universal method for uniquely establishing the phonemes of a language, and that some choices are effectively arbitrary, so you end up choosing one system over another on aesthetic grounds or on the basis of descriptive “simplicity” (unfortunately, there is no universal metric for “simplicity” in this context either.)

    If you’ve not come across it already, you may like this presentation, which goes into some detail on quite different phonological analyses of the same languages made by undoubtedly competent scholars, and leading up to the notion that Proto-Chadic may have been a language with no vowels whatsoever (yup, not even one, not even “syllabicity” or the like):

    https://www.eva.mpg.de/lingua/conference/08_springschool/pdf/course_materials/Wolff_Historical_Phonology.pdf

    (I can’t help but think, myself, that this is something of a reductio ad absurdum, reflecting the fact that the wrong questions are being asked. But Wolff knows vastly more about this than I do.)

  114. Stu Clayton says

    That implies Proto-Chadic had zero vowels (ZV). Not quite the same thing as no vowels, as I have learned here to my great dissatisfaction. Linguists and nature abhor a vacuum, is it not ?

    Even folk sayings exhibit this propensity. “If the shoe fits, wear it!”. The shoe is empty, so it must be filled with a foot.

    Another: “work expands so as to fill the time available for its completion”.

    Entropy is on a roll.

  115. David Eddyshaw says

    I think the idea is essentially that Proto-Chadic might not have divided segments into the categories of “consonants” and “vowels” at all. If you prefer, you could say it had no consonants. (But an awful lot of strange vowels, like /k/.)

    The trick is that the only vowel that disobligingly survives the wholesale (relatively uncontroversial) explaining-away process is /a/; if you then decide that this too is a consonant [basically, the same as the “glide” consonant that drasvi’s Arabic grammarians thought was at the end of Arabic /a:/, analysed as /aC/], then, Bingo! NO VOWELS!

  116. Stu Clayton says

    Then it would be clearer, albeit less exciting, to state that it had no distinction between consonants and vowels as we distinguish them today. Its speakers might have said of us, in their future, that we are hair-splitters.

    A distinction is a difference that makes a difference. If it doesn’t, it is useless.

  117. David Eddyshaw says

    I’m not getting excited until they can incorporate tone into the Grand Universal Smush as well.

  118. Stu Clayton says

    “Only connect.”

  119. 1-Lameen: I too have often heard “Arábic” instead of “Árabic” from the mouth of L2 anglophones. My hunch is that it is a case of overgeneralization: because of the stress alternation between “Árab” versus “Arábian” (The latter still alive and kicking in expressions such “The Arabian nights”, “The Arabian gulf”…) L2 speakers assume (not unreasonably, let’s be fair) that the stress alternation is triggered by the added syllable (something common enough in English), and thus deduce that “Arabic” must be stressed on its second syllable.

    2-David Eddyshaw: If we leave aside the loss of the definite article in some varieties of Romani spoken in regions where the dominant language was article-free (I am certain that this has been shown for Russian Romani), one possible candidate for prehistoric definite article loss is Berber: see this article (no pun intended)-

    https://journals.openedition.org/encyclopedieberbere/2602

    More broadly, Joseph Greenberg wrote an article (“How does a language acquire gender markers?”) on the topic (Basically, free demonstratives which used to exhibit agreement turn into bound gender markers via an intermediate article stage). In “Swahili and Sabaki” (pages 339-340) you will find some discussion of Greenberg’s thesis relating to the loss in a majority of Sabaki languages of the definite article, AKA “preprefix”.

  120. David Eddyshaw says

    Greenberg’s idea that preprefixes derive from a definite article seems to be historical speculation rather than something that can be seen in actual contemporary languages, but it may well be true, of course. I think there is good reason to imagine that the class affixes themselves derive from article-like clitics; in “Grassfields Bantu” Aghem, the class prefix is absent on the noun if it appears on an agreeing verb or if it appears on a following demonstrative, and the corresponding suffixes in Oti-Volta even now are loosely enough attached that suffixless forms appear constantly within NPs before adjectives and dependent pronouns, which themselves take the class-appropriate suffix instead. However, if you were going to try to attach an overall “meaning” to Oti-Volta class suffixes, it would have to be broader than “definite”; more like “not generic.”

    History is repeating itself: for example, in Gulmancéma, nouns typically take a class-agreeing proclitic when not used generically or preceded by a possessor: biga kuli “every child”, u biga “her child” but ki biga “the/a child”; and in Ditammari the cycle is complete: the proclitics have fused to the stem as class prefixes (a development which seems likely to be connected with the fact that in Ditammari there is a degree of phonological erosion of the inherited class suffixes severe enough that it would otherwise have resulted in frequent ambiguity.)

    The commonest way of losing definite articles seems to be that they are not actually dropped, but end up losing their specific meaning by being generalised to all contexts (with “generic” uses holding out against the process for longer than specific-but-indefinite.)

  121. David Eddyshaw: Fascinating. Your description of what is happening in Ditammari, where the proclitics have fused with the noun despite the old inherited morphology not having disappeared, reminds me of the rise of subject clitics in many Romance varieties, which took place despite the inherited suffixed person-markers being typically alive and kicking at the time.

    What you write about Gulmancéma and Ditammari does seem to indicate that Greenberg was basically correct. His article contains a great deal of data, so I would hesitate to call it “historical speculation”.

  122. David Eddyshaw says

    I’m pretty sure Greenberg was correct in his general thesis, at least as far as Niger-Congo goes. (I’m fairly sure I’ve read the paper itself, but can’t seem to locate it at present.) This is, after all, the Good Greenberg, on his home territory of typology, which he Owned. Respect!

    I just meant that I’m not aware of any Bantu languages where the preprefixes actually do function as definite articles at present (but that doesn’t mean a lot. I’m in no sense an expert therein. I would be extremely pleased to be pointed at an actual example.)

    The fusion of class proclitics to become class prefixes has happened also in a closer relative of Gulmancéma, Konkomba, where the erosion of the original suffixes is if anything even more drastic. There’s an interesting (if intermittently somewhat frustrating) grammar of Konkomba by a L1 speaker, Gbandi Adouna:

    https://tel.archives-ouvertes.fr/tel-00416375/document

    On the other hand, the Gurma language I’m most familiar with, Moba, which is even more closely related to Gulmancéma than Konkomba, has nothing like this at all, and neither does Naténi, which is the closest relative of Ditammari within Eastern Oti-Volta.

  123. David Marjanović says

    Like Arabien, arabisch gets second-syllable stress in German. Whether Araber gets stressed on the second or the first syllable seems to be a shibboleth for education.

    “…depending on what you make of the fact that personal-pronoun objects always precede the verb in Eastern Oti-Volta. …”

    As in Europe…

    In most of Europe they follow the verb by default: “I see you“.

  124. David Marjanović says

    I don’t know if “all language is MERGE” is falsifiable, but it certainly looks that way…

    That was a late-night version of “I don’t know if ‘all language is MERGE’ is unfalsifiable, but it certainly looks like it’s indeed unfalsifiable”.

    And indeed the new paper does not falsify Minimalism…

    Darwin later. 🙂

  125. By South Arabian, Al-Jallad does not (despite the confusing name) actually mean any form of Arabic. It’s a separate branch of Semitic …

    Arab: relating to the people or the culture.
    Arabic: relating to the language or the culture.
    Arabian: relating to the geographical region, i.e. the peninsula.

    <rant>

    Here’s why I’m confused: I don’t know any Arabi*, but I know enough Hebrew to see the similarities: the triliteral roots, Hebrew retaining letter Ayin despite its modern pronunciation being a glottal same as Alef, because historically Ayin denoted a pharyngeal fricative. (BTW could somebody put a rocket up wikipedia on ‘Hebrew Alphabet’! A glottal is not pronounced like the ‘tt’ in ‘butter’ in my English — but only in caricature London drawl — not even in Cockney.)

    To the point of the present discussion, Hebrew definite article ה arises from the posited *han- prefixed; but in Hebrew lost the trailing nasal, whereas in Arabic the h- reduced to glottal and the -n via an irregular sound change to -l, whereas other Arabi*s retained -n or assimilated -m or geminate -a-.

    *han- “emerged in the central Levant sometime in the late second millennium BCE, after the diversification of Northwest Semitic” [Al-Jallad]

    And yet Al-Jallad hardly mentions Hebrew — indeed in his ‘Manual’ there’s a family tree with root ‘Common Semitic’ and all sorts of Arabi*s including Canaanite, MSA, Ethiopian, Babylonian, but no Hebrew. Whaʔ? Not all of those count as Arabian peninsula, by @Y’s rubric; whereas Hebrew couldn’t get more central Levant.

    So I had been presuming that the various (South/North) Arabians were a lot closer to Arabic than Hebrew is to Arabic. Is that not the case? (Hebrew seems to belong with Aramaic in ‘NorthWest Semitic’ which supposedly diversified before the arrival of *han-.)

    Or is Al-Jallad just pretending Hebrew doesn’t exist? Because Religion (presumably).

    Is anybody brave/foolish enough to give a metric of how remote (Modern) South-Arabian is from Arabic compared to within a family I know like Romance or Germanic?

    I suppose it’s futile to ask why Sudanese is pretty close to standard Arabic, whereas next-door Ethiopian (Semitic languages) are as various as the South Arabians.

    </rant>

  126. David Eddyshaw says

    Or is Al-Jallad just pretending Hebrew doesn’t exist? Because Religion (presumably).

    No, Al-Jallad is completely innocent in this matter.

    Hebrew comes under “Canaanite” (whatever you may have gathered from the BIble.) It’s actually called “the language of Canaan” in the earlier strata of the Bible itself.

    This is standard usage in comparative Semitic studies, and it’s linguistically perfectly reasonable. Moabite, Ammonite and even Phoenician were very closely related to Hebrew. Ugaritic is also very similar (a fact partly obscured by its rather earlier attestation) and has been regarded as “Canaanite” too by modern scholars sometimes, although the actual people of Ugarit did not regard themselves themselves as Canaanites. Aramaic is distinguishable from Canaanite from about the beginning of the first millennium.

    I suppose it’s futile to ask why Sudanese is pretty close to standard Arabic, whereas next-door Ethiopian (Semitic languages) are as various as the South Arabians.

    Not futile, no. It’s hardly a secret that Arabic got to that part of the world rather later than Ethiosemitic. It hasn’t had as long to differentiate (and hasn’t been subject to anything like as much Cushitic substrate influence.)

    So I had been presuming that the various (South/North) Arabians were a lot closer to Arabic than Hebrew is to Arabic

    No, Arabic is closer (genetically) to Hebrew (and the rest of Canaanite) than to Modern South Arabian; it shares with Canananite/Aramaic/Ugaritic a definite non-trivial common innovation which MSA and Ethiosemitic lack, the yaqtulu imperfect.

    Old South Arabian is more complicated. At least one of the languages within OSA also has the yaqtulu imperfect; nevertheless, OSA doesn’t seem to form an actual subgroup with Arabic (but I think a lot of work yet remains to be done to sort all this out definitively.)

  127. David Eddyshaw says

    A glottal is not pronounced like the ‘tt’ in ‘butter’ in my English — but only in caricature London drawl — not even in Cockney.

    You’d be surprised. Self reflection is not always a good guide in these matters, especially if that takes the form of saying words over to yourself slowly. Did you know that syllable-final voiceless stops are preglottalised in UK English? (I can’t say I would have noticed this myself, left to my own devices.) The “glottal stop” version of /t/ happens just because the actual alveolar closure frequently gets lost in rapid speech. Speakers who aren’t in the least Cockney very often do this too: they can’t hear it, just as most normal (i.e. non-Hatter) English speakers can’t hear the difference between the /t/ in “stop” and the /t/ in “top.” If you try to hear it by listening to yourself, the very fact that you are doing so alters your pronunciation. Heisenberg …

    [Americans, of course, pronounce the /t/ in “butter” as [d], but they haven’t had our advantages.]

  128. Self reflection is not always a good guide in these matters, …

    Fair comment. By ‘my English’ I more meant listening to my brothers/sisters’ English (outer West London) to their partners and the partners’ brothers/sisters (variously outer South London: closer to a glottal in ‘butter’, posh North London: no glottal, very outer North-West London/practically in the country: not at all uncouth). I wouldn’t trust listening to myself alone, not only for the reasons you give, but also because I’ve lived outside London/outside Blighty for decades.

    (To correct my post: the wp uses ‘button’, also the ‘p’ in clipboard; but syllable-final is not where Alef/Ayin are problematic for English speakers.)

    If you’re trying to explain Hebrew glottals, surely pick other exemplars! (As you say, it doesn’t help Americans at all.) You should be picking words where the Alef/Ayin sound appears word-initial: ‘a apple’, ‘I’ll be …’, ‘agriculture’.

    I used to practice waiting at the bus stop outside the kibbutz, by pronouncing the bus company’s sign that featured Alef as its logo (אֶגֶד) — and reading left-to-right could fancifully be construed as ‘taxi’.

  129. David Eddyshaw says

    I agree that “butter”* is unlikely to help anybody who doesn’t already know what a glottal stop sounds like. It’s part of a sort of entrenched tradition that turns up in the old “Teach Yourself” books (good as those often are.) Copied from one book to another …

    Even trained linguists are often not too good at this. Words like Kusaal sʋ’ʋg “knife” are described in all the accounts of Kusaal phonology I’ve seen as being pronounced with [ʔ]: [sʊʔʊg]. But they’re not …

    I believe that modern Israelis have neatly solved the problem by not actually pronouncing alef at all. (Or he, for that matter.) But there are those among us who will Actually Know.

    * “Bottle” is actually better. But still not very helpful to someone who doesn’t already know the answer before your tell them.

  130. I definitely hear some Americans use a glottal stop in kitten and mitten. Not in bitter or better. Someone surely has picked at that already.

    Modern Hebrew has merged, in the speech of many, /h/ into /ʔ/. Not mine, though.

  131. I agree that “butter”* is unlikely to help anybody who doesn’t already know what a glottal stop sounds like.

    A better example for modern English speakers is probably “uh-oh”. Word-initial doesn’t work so well because phonemically there’s nothing there: we say “an apple”, not “a apple” (well, most of us anyway.)

    Almost all Arabic colloquials have historically lost the glottal stop, but quite a few have regained it by changing q to ‘, and sometimes also by reborrowing from Standard.

  132. The regional New Haven area pronunciation of „Connecticut“ has two glottal stops, which I only consciously realized upon hearing a Yalie imitating some local women.

  133. I could have sworn I’d already told this joke (I think I heard it told about Scotspersons) at LH, but apparently not:

    Son says “Pass the bu–er.” Father: “Don’t say bu–er, say butter” (carefully enunciating the /t/). Son: “Pass the butter.” Father: “Tha–s be–er!”

  134. @DM, I did not mean that “most” European langauges have it, just that it is familar from French, Italian etc. (je t’aime).

    Pre-verbal position of pronouns in Russian is at least consistent with functional and prosodical considerations.

    я тебя вижу “I you see” usually means “you are visible”.
    я вижу шпиль над горизонтом “I see spire over horizon” informs of a new object entering my field of sight.

  135. я вижу шпиль над горизонтом

    Sounds like the start of a Blok poem.

  136. “uh-oh”

    [ˈʌ̆ʔ˦oʊ˨] says Wiktionary. Uh-oh…I personally used as an example of the glottal stop in Russian mm-mm ( [ˈ(ʔ)ḿ̩.ʔm̩̀] says Wiktionary about the English version) and uh-uh [ˈ(ʔ)ʌ̃˧.ʔʌ̃˩].

  137. [ˈʌ̆ʔ˦oʊ˨]

    I often say it (and hear it?) with an initial ɦ.

  138. @Lameen: quite a few have regained it by changing q to ’

    What’s the distribution of that phenomenon? I’ve heard it described as urban Palestinian, and it exists in [ʔal ʔuds], but what other cities? WP says that this sound change also occurs in Syria and Lebanon; is there a similar urban/rural divide there as well?

  139. David Marjanović says

    not even in Cockney

    I’ve probably never heard the genuine article. But, well outside Cockney, I wouldn’t understand half the words this bloke says if I didn’t know to force myself to interpret almost every one of his pauses as /t/.

    Lots of Americans do this before nasals, e.g. in button or bottom, but nowhere else.

    Did you know that syllable-final voiceless stops are preglottalised in UK English? (I can’t say I would have noticed this myself, left to my own devices.)

    There are people who do it so heavily that when they speak through bad microphones, I get trouble breathing just out of empathy.

    most normal (i.e. non-Hatter) English speakers can’t hear the difference between the /t/ in “stop” and the /t/ in “top.”

    Oh, that’s only after they’ve been reading & writing for long enough. Little children routinely go for D when trying to write star and for G when trying to write sky, for example. And to my ear, most native speakers really go all the way to the voiceless lenis there phonetics-wise.

    pronounce the /t/ in “butter” as [d]

    Well, it’s not a plosive, it’s a (the?) flap: rather than holding the tongue there and then pulling it away, they flick it. And then they hear a Spanish r as the same thing, which it isn’t (despite usual IPA practice!), but it can sound pretty similar.

    It’s part of a sort of entrenched tradition that turns up in the old “Teach Yourself” books (good as those often are.) Copied from one book to another …

    We have those in German, too. [z]: “s, like in Rose.” So, [s] then. *Merkel eyeroll* This one works everywhere north of the White-Sausage Equator, but south of it it fails. The only trick that would work everywhere would be “like a bee: bsssss…” – and that would also take care of the actually voiced [b], which has an even more restricted distribution!

    A better example for modern English speakers is probably “uh-oh”.

    …which I’ve seen spelled ut-oh several times.

    we say “an apple”, not “a apple” (well, most of us anyway.)

    “A apple”, with a glottal stop inserted into the vowel cluster, seems to be an AAVE thing. Search for Chris Rock on YouTube.

    What’s the distribution of that phenomenon?

    Also Egyptian and Maltese.

    (among, probably, others)

  140. reminded
    – [ʔ] is used among Palestinian boys (talking to each other)
    – [g] is used among Jordanian boys
    – [ʔ] is used by both groups when talking to girls
    – [g] is used by both groups in mixed (Jordanian and Palestinian) boy groups

    here p 9.

  141. half the words this bloke says if I didn’t know to force myself to interpret almost every one of his pauses as /t/.

    No Marcus Bronzy isn’t Cockney. There’s strong elements of (I think) Caribbean Vernacular English — I’m guessing North London around Tottenham (to’nim), but my ear isn’t in these days.

    A fair proportion of his pauses aren’t /t/. In particular listen out for ‘in pa’i’ular’ — at about 0:20 here.

  142. PlasticPaddy says

    Di’ ‘a’e’s a lo’ w’v blokes wha’ ‘as no ‘eef o wha’ ‘s i’ lo’sa figh’s..

  143. I’ve heard decidedly white Americans say “A apple”.

  144. @Y I have also, and there’s usually a noticeable glottal component between the indefinite article and the noun. EDIT: four speakers: Tucson, Philadelphia, Chicago and Baton Rouge.

  145. David Eddyshaw says

    I have always supposed that the trick to sounding like a Londoner is to omit all consonants, or at least all stops (other than glottal, of course. That would lead to incomprehensibility. Obviously.)

    My Hispanic son tells me that this works for Andalusian Spanish too, and without any need for glottal stops, either.

  146. … Cockney. I’ve probably never heard the genuine article.

    FFS I can’t find a single youtube vid with yer akshull Cockney. Not Michael Caine: he’s South London. Not Adele: she’s Tottenham (not Caribbean)/Essex/South London. Not James Corden he’s West London — even further West than my accent.

    No it’s not rhyming slang, nor Kenneth Williams — as brilliant an ear as he has, and he could be close. Barbara Windsor in the Carry On films is the best I could find. They both get increasingly Cockney from about 1:00 on here — but they’re maybe ‘putting it on’. There’s a nice ‘sending me u’ gu’liss’ from Barbara about 2:10.

    There’s heaps of vids alleging they’re explaining ‘Cockney’ but what they mean is some generic London drawl. THERE’S NOT _A_ LONDON ACCENT. Dropping your /t/s and /h/s does not make you Cockney nor even a Londoner.

  147. IIRC Cockney got extinct in about in the ’70s or ’80s, if not earlier?

  148. the trick to sounding like a Londoner is to omit all consonants

    Corky St. Clair (i.e. Christopher Guest), in Waiting for Guffman, figured that out to perfection. “’Ow ’ah ’oo ’oo’ay?”

  149. AntC, how’s this:
    https://www.ivoox.com/en/fred-bason-s-diary-episode-1-audios-mp3_rf_78080595_1.html

    That is the diary of Fred Bason, a book dealer and raconteur and Cockney to the bones. Bason died in 1973, and I don’t know if recordings of his exist. The diary is read by Clive Farahar, a rare book dealer who grew up in London. Farahar is, I think, not at all Cockney; how good is his rendering?

  150. AntC : Michael Caine can not pronounce consonants other than glottal ones. That is an established fact.

  151. David Eddyshaw says

    @AntC:

    YouTube tells me that ITV_plc has blocked me from seeing this because of my country of origin. I am astonished that it should be even legal to deny a proud Britisher the right to see YouTube clips of Carry On films in his own home! I didn’t vote for Brexit for this!* Take back the remote control!

    * Or at all. But let’s not split hairs over this.

    [Ah. I see that it is a mere interview, rather than a clip from the Classics of British Cinema. Stll, harrumph!]

  152. David Eddyshaw says

    Yes, thanks, V. The apoplexy has now passed, without serious sequelae. I am now in a pleasant fug of mono no aware “fings ain’t wot vey used ter be”, and my blood pressure has returned to normal.

    [V’s kind enquiry after my health has mysteriously disappeared, but I will leave this comment as a reassurance to the many Hatters who doubtless shared the concern.]

  153. I’m glad you’re at a better place mentally. (I thought you were joking when you said that you thought that could not happen to you).

  154. Lots of Americans do this before nasals, e.g. in button or bottom, but nowhere else

    Not before /m/, only /n/, e.g. kitten [kʰɪt̚n̩] ~ [kʰɪʔn̩]. That is or used to be the standard pronunciation, but the “flap /t/ between vowels” rule is being generalized into the pre-/n/ environment, so that Kids These Days (at least on the West Coast) say [kʰɪɾә̃n].

    For Cockney, ‘ow about Peter Sellers?

  155. I am reminded of a science-fiction short story I read a long time ago, involving an alternate universe where a united Roman Empire was re-created a few centuries after the fall of the Western Roman Empire, and then went on to colonize much of the planet (including Eastern North America, where the story is set), with China its main global rival. Its technology is basically medieval, but it has a kind of computer technology whose foundation is “Pāṇinian” formal logic.

    Uchronia does not have all alternate universes, but it does have quite a few. Perhaps seeing others will jog your memory?

    http://uchronia.com/bib.cgi/diverge.html?o=250

  156. David Eddyshaw says

    My sister, though a mere six years younger than me, actually releases the /t/ in “kitten.” Young people of today …

  157. @ Owlimirror : Either Paul Anderson, Roma Eterna, or The Years of Rice and Salt by Kim Stanley Robinson. EDIT : None of those two compute. Ideas?

  158. I have always supposed that the trick to sounding like a Londoner is to omit all consonants, or at least all stops

    I thought that was the trick to sounding like a Dane.

  159. Lars Mathiesen says

    You called?

    Actually, that’s the Jutes. “A æ å æ ø i æ ø æ i å å æ ø i æ å?”

  160. I thought the stereotype was that Danish has very few phonemic vowels? Although in Lars’ example I see only five, so that might not be contradictory. Compared to the rest of Scandinavia, that’s few.

  161. A friend of mine complained that English freinds of his bilinguial (English-Russian) wife kvakayut (qua’a’a is the sound frogs make).

  162. (But an awful lot of strange vowels, like /k/.) – That’s what Elmedlaoui’s Berber is supposed to have.

    In the article that I used to illustrate “alif” Arab ideas are compared to the Optimality Theory. But here the relationship can be more than just similarity: I think the OT was inspired by Dell and Elmedlaoui’s work, and Elmedlaoui must be familiar with Arab phonological ideas.

  163. Roma Eterna is Silverberg. It does not look like what Etienne meant, but its WP article (Roma Eterna) has a See Also.

  164. > Roma Eterna is Silverberg

    It is indeed. And I’ve read it! (a long time ago) Seemed quite strange, even at my tender age, even in translation. I haven’t read The Years of Rice and Salt, but I’ve read bad reviews of it. And those are novels, not short stories in the first place.

    I did like some Silverberg, though, like Lord Valentine’s Castle. I don’t know if it will be visited by the suck fairy if I read it again as an adult.

  165. I haven’t read The Years of Rice and Salt, but I’ve read bad reviews of it.

    I find Robinson somewhat frienghtening. I only saw some of his recent novels, he clearly reads Wikipedia and borrows thousands factoids, terms and concepts from there, sometimes even adding sub-chapters wholly dedicated to piled-up undigested terms and concepts. If there is a need to do so, he could at least add pictures, hyperlinks and footnotes…
    The Years of Rice and Salt must be less scary (it was written in pre-WP times).

  166. Years of Rice and Salt was an unusual book – framed as the repeated reincarnations of a couple of characters whose lives range from India to Andalus to San Francisco Marin, in a world where the Black Death took out Europe (but, for some reason, not North Africa or Anatolia or…) Not one of his best in my view; it had some great moments, but Robinson does not understand non-American cultures nearly as well as he thinks he does, which for a world like this one is a big problem. Try Escape from Kathmandu instead.

  167. “Roman Empire” and “New World” reminded me of something . . . Somtow Sucharitkul’s (S. P. Somtow) Aquilia stories. But looking at those suggest important differences in details, e.g. the Roman Empire never fell (The Uchronia website has the divergence date of 200 BCE), not that it rose again.

    @Etienne: I don’t suppose you recall whether the Empire was Christian or not?

  168. @dravsi : Well Wikipedia started being a thing people talked about around 2003, IIRC, and the book was published in 2002, so just barely.

  169. I can’t help but think, myself, that this is something of a reductio ad absurdum, reflecting the fact that the wrong questions are being asked.

    It seems pretty easy to diagnose the wrong questions: 1) sound change operates on allophones, not phonemes, and hence the comparative method (= inverse inference of sound change) must too; 2) “underlying structure” is basically internal reconstruction as re-invented by people illiterate in historical linguistics; 3) therefore any introduction of “underlying structure” into historical reconstruction is almost a guarantee of being hopelessly confused about relative chronology. For reconstruction purposes (perhaps more generally too), it is useless to claim that e.g. an [u] “is /w/” (or that /u/ “is //w//”) without also claiming that it actually was a non-syllabic [w] at some point.

    The deepest problem is thus clearly the Semitic-derived or indeed traditional-Semitic-grammar-derived axiom “roots must be consonantal”. Something like [uma] can be √wm-a only if we have reason to think there was a pre-stage where this was something like *wVm-a or *Vwm-a. Otherwise its root should be considered to be √um-. Root structure is descriptive, not prescriptive.

    I like what the Chadicists have been doing with topics like vowel epenthesis but I hope they get a better grasp of realism in phonology sooner or later.

  170. David Eddyshaw says

    Yimas is another language in which you can (nearly) eliminate all vowel phonemes but /a/, partly by cheating with fronting and rounding suprasegmental prosodies, just as the Chadicists are wont to do, but nonetheless in an interesting way.

    Foley’s grammar goes into this in great detail; he unfortunately adopts a working orthography based on this analysis which is pretty opaque as far as the vowels are concerned, even if you’ve worked hard at absorbing his phonological stuff, with words like tkt “chair”, mml “Javan file snake”, and phrases like tnkntkntrm tktntrm amanantrm “my two heavy chairs.”

  171. OK, completely out of context, Mohamed Salah clearly just comitted an offense foul against the Moroccan goalkeeper in the penalty area during extra time. That should have resulted in a yellow card. The referee was clearly in favour of the Egyptians.

  172. David Eddyshaw says

    He evidently believes in spoiling the Egyptians.

  173. David Marjanović says

    framed as the repeated reincarnations of a couple of characters whose lives range from

    Oh, that’s a favorite method of the Disney franchise.

    (Except they’re called “descendants” instead of “reincarnations”. The family relations stay the same, though: the uncle is descended from the uncle, the nephews from the nephews. In the long run, even reproduction by nepotism isn’t asexual enough.)

  174. As I was exporing Wikipedia, I came across a description ; “The book is set in an alternate present in which Islamic countries form a prosperous, democratic and progressive First World, while underdeveloped Christian countries suffer from religious fanaticism“.
    An evil smile. I am the Second World, I remain invariant to such transformations:-E.

  175. @ David Eddishaw:

    > He evidently believes in spoiling the Egyptians.

    I don’t get that reference. Is it from the King James’ Bible?

  176. David Eddyshaw says

    Yup. Exodus 12:36.

    However, it’s a standard expression (or perhaps, rather, used to be a standard expression … tempora mutantur, nos et mutamur in illis …)

  177. WP says that this sound change also occurs in Syria and Lebanon; is there a similar urban/rural divide there as well?
    At least for Lebanese, it’s considered a standard feature, also usual in the Lebanese used in songs and films. I wouldn’t know whether there are rural dialects that don’t have it.

  178. From the introduction to my (excellent) A Dictionary of Syrian Arabic: English-Arabic by Karl Stowasser and Moukhtar Ani (where “Syrian Arabic” includes Lebanese, Jordanian, and Palestinian):

    Although ق has merged with ء in urban Syrian pronunciation, the two letters are distinguished in our transcription (q = ق, ʔ = ء) as an aid in word-recognition for those who are familiar with literary Arabic or with other Arabic dialects. In certain classicisms, however, the true ق-sound is normally used in colloquial speech; this sound is represented here by the symbol ; maq̈aal ‘article’, ḥuq̈uuq̈ ‘law’ (but this is also heard as ḥquuq, i.e. pronounced ḥʔuuʔ).

  179. > However, it’s a standard expression (or perhaps, rather, used to be a standard expression)

    Among people whose first language is English, and have read the Bible in the KJV version or in Vulgar Latin, rather than in Old Bulgarian or in Koine Greek?

  180. David Eddyshaw says

    Among people whose first language is English

    Well, yes. At least I have certainly encountered it in the wild, used without any explicit reference to the Bible.

    But I’m now in the uncomfortable position of someone retrospectively trying to explain a joke that wasn’t very good in the first place.

    I should point out, however, that having actually read the Bible in the KJV is by no means necessary for many of its locutions to be familiar. They’ve become common property (to the point that many speakers actually are unaware of their origin.)

  181. Yup. A representative example of usage:

    I wonder how many people in McMinnville know that when the church people bring their kids to your door, in those pressed red and white blouse and skirts, or blue slacks and a white shirt with a red sweater vest, neatly dressed and irresistible smiles, to get sponsors for their school’s Bike-A-Thon fundraiser, they privately call this ‘Spoiling the Egyptians’?

  182. David Eddyshaw says

    I’m actually familiar with it in deeply secular contexts, despite my own background (stealing paperclips from work would be the first thing that comes to mind for me. Not that I ever would, myself, you understand …)

    However, idly googling the phrase, it seems to me that knowledge of it is dying (if not extinct) among those unfamiliar with the original source. O tempora! O mores! I feel that apoplexy coming on again …

  183. @David Eddyshaw: Whereas, in the old days to learn the phrase, just exposure to mainstream British culture, It would have been enough?

  184. David Eddyshaw : “I feel that apoplexy coming on again” — don’t you’re fine. I was trying to explain something complicated, and now I don’t have to. O tempora! O mores! — I thought everyone with an education knew that phrase.

  185. About /q/

    There is Bergsträsser’s Sprachatlas von Syrien und Palästina .

    The map is p. 78 of the pdf.

  186. He evidently believes in spoiling the Egyptians.

    I didn’t get that one, not being that familiar with the Bible.

    Having checked the passage, I’ve belated realised that it should be understood more along the lines of despoiling the Egyptians.

    Meanings nicely summed up here:

    https://biblehub.com/exodus/12-36.htm

  187. § 10. 3) [ق] :6/13/14/15/19/39/41/42/50/51/52/53/55/58/59/61/63/67/78/83/85/88/90 8/60. 15/45. 26. 26/38. 26. 33. 37. 45. 53/64. 68/68/79/79.82. 85. — m8 S. 116. 203. 388. 394. 450 (nur die Beispiele für ǧ).
    Karte 4 (Tafel XXIV). — Die Aussprache ʾ herrscht außer in einem besonderen Gebiet in den meisten Städten; doch sprechen auch in den Städten teilweise die Ungebildeten (50. 45) bezw. g (16) [1], wie umgekehrt auch auf dem Lande die Gebildeten bezw. die Christen gelegentlich ʾ sprechen (s S. 1; so ist wohl auch das von mir 57 gehörte ʾ aufzufassen). g wird außer in einem besonderen Gebiet auch von allen Beduinen gesprochen, und zwar meist mit sehr weit hinten liegender Artikulationsstelle. Bei einem Teil der Beduinen ist ق in manchen Wörtern ǧ geworden (s. u.); wird gewöhnlich wenig energisch artikuliert; k gleicht sonstigem ك. — Im allgemeinen stimmen meine Ergebnisse mit den Angaben von Littmann S. 6 überein.

    Anm. Christie schwankt c1. c2 zwischen ʾ und , c8 zwischen , k und g; im letzteren Falle ist wohl k gemeint. — Landberg S. 329 betrachtet ʾ als Sprachfehler und schreibt daher, auch wo er es gehört hat, . — Littmann a. a. O. gibt für Aleppo an, daß nur die Christen ʾ sprechen, die Muslims dagegen ; ich habe von einer gebildeten Jüdin ʾ und von einem ungebildeten Muhammedaner gehört. Pourrière hat nur ʾ. — Musil, schreibt ž für ǧ. Für m5 finde ich bei ihm einen Beleg für ǧ (172); bei der großen Menge des Materials, das er aus el-Kerak mitteilt, kann mich dieser Beleg nicht überzeugen, daß gegen alle Erwartung der Übergang dort tatsächlich stattgefunden hat.

    [1]) Auch 27 gehört das von mir notierte k vielleicht nur den Ungebildeten an.

    (p. 19(30))

  188. David Eddyshaw says

    I knew it would turn out to be simple in the end.

  189. Ouch.
    They found q~ʔ in Safaitic.

  190. They found q~ʔ in Safaitic.

    So Al-Jallad argues, but the evidence is rather slim and not particularly convincing.

  191. Trond Engen says

    David E.: The commonest way of losing definite articles seems to be that they are not actually dropped, but end up losing their specific meaning by being generalised to all contexts (with “generic” uses holding out against the process for longer than specific-but-indefinite.)

    Is Arabic going that way? Or would it, if Classical Arabic hadn”t frozen formal grammar in time? I’ve been thinking that, or at least that the definite is the unmarked form, but it may be an impression based on nothibg but the density of definite articles in any Arabic phrase.

  192. David Eddyshaw says

    A lot of Arabic loanwords in West African languages incorporate the article as part of the stem. Mind you, so do French loanwords (the single most widespread French loan in WOV is without doubt lampo “tax”; even Dagbani has it.)

  193. And Spanish tended to incorporate Arabic nouns with articles in borrowings, of course.
    Question: where you have both definite and indefinite articles, does this tend to lead to a situation where most nouns have an article, but the balance between them becomes relatively stable? It strikes me that in some languages where forms with the definite article have marginalized anarthrous forms, there was only a definite article when this happened. But counter-examples would be welcome.

  194. David Marjanović says

    All Arabic nouns borrowed into Spanish seem to have the article still on (but not those borrowed into Italian). French nouns all the way to Chinook Jargon tend to keep their articles on as well.

    I knew it would turn out to be simple in the end.

    The cities turned /q/ into [ʔ]; the Bedouins turned /q/ into [ɢ], sometimes further into [g] or even [dʒ]; meanwhile, /dʒ/ turns into [ʒ]. All the while, people are aware of Classical [q]; sometimes they try to pronounce it, fail, and use [kʰ] instead. Let that stew for a thousand years and revel in the general confusion.

  195. All Arabic nouns borrowed into Spanish seem to have the article still on

    Many, but not all: see, for example, loco (an adjective, but frequently nominalised), dado or mohín(o)

  196. where you have both definite and indefinite articles, does this tend to lead to a situation where most nouns have an article, but the balance between them becomes relatively stable?

    Judging by North African Arabic and Berber under French influence, nouns only get borrowed with the definite article, almost never the indefinite one. (The only exception I can think of is diblat from “des blettes” for “chard” in some dialects, which fits its mass noun status.)

    Is Arabic going that way? Or would it, if Classical Arabic hadn”t frozen formal grammar in time? I’ve been thinking that, or at least that the definite is the unmarked form

    That has actually been argued for southern Moroccan Arabic (in Turner 2013), but outside that region the definite article still looks very stable.

  197. David Eddyshaw says

    No SF novel called The Years of Rice and Salt could possibly be any good. It’s like calling your son “Plantagenet” and expecting him to grow up be a credit to the family. I’m sorry, but there it is.

  198. Stu Clayton says

    Sounds like a novel by Pearl S. Buck.

    The young’uns may have no idea what I’m getting at. For the older and more clubable, sat nomen.

  199. It’s like calling your son “Plantagenet” and expecting him to grow up be a credit to the family. I’m sorry, but there it is.

    Are you casting aspersions on the Duke of Omnium? Planty may have been a bit barmy about decimal coinage, but surely he was a credit to his family, the peerage in general, and (God save the Queen!) his blessed isle.

  200. David Eddyshaw says

    Actually, I had the loathsome Woodrow Wyatt in mind …
    (… though I know no ill of Pericles Plantagenet, whom I feel ought to be cut considerable slack. His sister has also made some bad choices.)

  201. Sami Jiries notes that *q > ʔ is already recorded for the Palestinian pronunciation of Syriac in the 13th century, presumably carried over from their native Arabic pronunciation (and, as Lameen notes further down, the folks living there were already referred to as palesṭīnāyē).

  202. ” All the while, people are aware of Classical [q];…”

    I thought “simple” referred spoiling the Egyptians, but the wonderful thing about Arabic isoglosses and sociolinguistical variable is that they are known to be variables to native speakers.

    When some people in a locality say it this way, and some say it that way, and children when they learn to speak say it some other way, of course people have opinions. And you have situations like that described by Enam Al-Wer for Amman above, when “boys switch to glottal stops in presence of girls”.

  203. @Y, O! Thank you. There is an old (1958) article Remarks on the Historical Phonology of an East Mediterranean Arabic Dialect by Irene Garbell, it has a passage:

    3. Stage 3. (11th-15th centuries)
    ….
    A further significant change took place at that stage, viz. the merger of /q/ with /ʔ/, which is witnessed for the contemporary pronunciation of Syriac in Palestine by Barhebraeus.[41] This development, which was doubtless preceded by a ‘de-emphatization’ of /q/ (for otherwise the latter would have merged with /ʕ/), possibly represented at first a local peculiarity, but in the course of time spread to other urban and later also to most rural communities of the Eastern Mediterranean region. [42] It was favored by the progressing weakening of /ʔ/; following the merger this phoneme could now in all positions, including word final and syllable final.

    see
    I wanted to take a look at what Barhebraeus wrote, but realized that I am not sufficiently motivated…

  204. Also see a note about /q/ in the Levant by Jérôme Lentin here.

  205. David Eddyshaw says

    No teenage boy wants to sound like a hiq when there are girls about.

  206. It’s like calling your son “Plantagenet” and expecting him to grow up be a credit to the family.

    Thinking of Charles Montgomery Plantagenet Schicklgruber “Monty” Burns?

  207. Thanks, drasvi!
    I wonder if the other sound change mentioned by Bar Hebraeus, *θ > f, is attested in any modern variety.

  208. Cockney Arabic!

  209. θ > f
    These two are what I confuse. Arabic articulation of /f/ is somewhat different from Russian and sometimes I mishear it (as articulated by Arabic speakers) as θ. θ is a difficult sound for me as a Russian, literally “difficult”: a simple sound, but my tongue gets tired of making it. Nevertheless I confuse them in this direction. My Arabic-speaking friend immediately recognized my problem when I complained, but I do not remember why it was familiar.

  210. I sometimes wonder why Russian got θ > f while Bulgarian got θ > t, In Greek loans.

  211. David Eddyshaw says

    The Kusaal Bible usually renders English /θ/ as t, including word-finally, but for unclear reasons features a book about Ruf the Moabitess, and also mentions a tall fellow called Goliyaf from Gaf.

    Actually, Kusaal /s/ is quite often realised [θ] by some speakers, at least in some contexts. I’m not sure why the Bible translators decided to go with /t/. I can’t actually think of any common nouns borrowed from English words with /θ/, so I don’t know what the “natural” equivalent would be. None of the other languages that Kusaal has borrowed from include this exotic phoneme*, so no help there either.

    [Looking through Naden’s dictionary to see if he had any loans from English vocabulary featuring /θ/, I discovered that pʋlis “police” has acquired a perfectly-regularly-corresponding singular pʋlʋŋ “police officer.” I Did Not Know That. It makes me happy.]

    * It turns up as t in words ultimately from Classical Arabic containing /θ/, like Talaata (daar) “Tuesday”, but all such words seem to have got mediated via other languages on the way to Kusaal.

  212. It makes me happy too, and I’m not even a Kusaalist.

  213. Maybe for Russian priests Ф was “that strange Greek sound” (/f/ was absent in Slavic and there is a regional pronunciation /xv~xw/) and θ was “that strange Greek sound”.

  214. David Marjanović says

    Yes; the question is why that didn’t happen in the Balkans (Serbian too went with /t/ for θ).

    But to return to glottal stops for a minute… I keep forgetting to mention that there’s a Standard German accent (used in some sort of Rhineland or so) where glottal stops are phonemic. First, -pen, -ten, -ken come out as [ʔm̩ ʔn̩ ʔŋ̩], so we get a /ʔ/ and a /ŋ/ at a bargain price. Second, -den also becomes [ʔn̩], but this time, as far as I’ve noticed, the [ʔ] is lenis, while it’s fortis in the other three, so apparently we have to set up two different /ʔ/!

    (As I recently complained elsewhere, whenever you come across a presentation of a sound system and the claim “this is the sound system of Standard German”, every word of that claim is wrong, even if you count sound and system as separate words.)

  215. That’s my theory also: Russia was more distant from Greece, and they had little contact with actual Greek speakers and for them Ф looked visually similar to θ.

    And that /f/ was absent in Slavic, and perceived as foreign.

  216. I wonder if the other sound change mentioned by Bar Hebraeus, *θ > f, is attested in any modern variety.

    Yes, it’s found in some Anatolian Arabic varieties – the Siirt group, apparently. Very much not mainstream though.

    And you have situations like that described by Enam Al-Wer for Amman above, when “boys switch to glottal stops in presence of girls”.

    There is one town in Algeria, Tlemcen, whose dialect traditionally changes q to a glottal stop. Nowadays, apparently, women still do this while men and boys sedulously avoid it for fear of sounding effeminate.

  217. David Marjanović : Do you think that the German speakers I interacted with while in Cologne (mostly around the university and around Ehrenfeld) were not speaking Kölnisch? They seemed to speak standard German. (in my limited capacity to distinguish)

  218. David Marjanović says

    They could have been speaking this; note in particular the section on the intermediate position.

  219. i went looking for other athenians to play with p. p. [etcetera] wyatt, and what did i find but an a. alcibiades hohenzollern. just thought i’d share the joy, though i’ve got no idea how 16thC franconian fits into any of this germanistik.

  220. Albert Alcibiades, Margrave of Brandenburg-Kulmbach! A splendid name, and a splendid beard.

  221. …and our metalheads would aprreciate the costume.

  222. [ʔŋ̩] … so we get a /ʔ/ and a /ŋ/ at a bargain price

    I can undercut that bargain: Ng (pronounced [ŋ̍];

    caused enormous difficulty for our English-speaking receptionists taking messages from Mr. EnGee.

    “is a Cantonese transliteration of the Chinese surnames 吳/吴 (Mandarin Wú) and 伍 (Mandarin Wǔ). ”

    Now under what sense of ‘dialect’/Fāngyán/方言 is Wú a _transliteration_ of Ng?

    (The Mr. Engee in question — Malaysian (Straits) Chinese, immigrated to Britain — also had a single-syllable personal name; and managed to get a UK personal number plate at a bargain price, because DVLC thought it just a random scrabble of letters.)

  223. David Marjanović says

    managed to get a UK personal number plate at a bargain price, because DVLC thought it just a random scrabble of letters.

    Awesome!

    Now under what sense of ‘dialect’/Fāngyán/方言 is Wú a _transliteration_ of Ng?

    None. It’s the Mandarin cognate. (Or, in traditional Chinese terms, the Mandarin reading of the character 吳/吴.)

    I can undercut that bargain:

    Sure, but 1) that doesn’t give you a /ʔ/, and 2) German isn’t supposed to have either a /ʔ/ or a /ŋ/, and here we seem to get both.

    (In mainstream English you have to set up a separate /ŋ/ because of singer vs. finger. Such a difference does not occur anywhere in German that I know of, so you can get away with declaring every [ŋ] to be a /n/ that is followed by a /k/ or an otherwise disappeared /g/ or, if word-final and syllabic, preceded by one – but that doesn’t work in the accent I described.)

  224. PlasticPaddy says

    @dm
    re ng with hard g in German, there is no barrier to producing n+hard g in adjacent syllables. This applies both to borrowings likeTangente and to native words like Wohngemeinschaft. But I suppose you distinguish n+hard g in these cases from a monosyllabic ng, even though that distinction is not always clear to me (ok, potentially the speaker could take a breath between the n and the hard g, without extending the n…).

  225. or an otherwise disappeared /g/
    That would be an underlying phoneme and therefore cheating. As long as there is a contrast between Tann /tan/ and Tang /taŋ/, /ŋ/ is a phoneme, even if a positionally restricted one.
    The analyses of Standard German pronunciation I know all have /ŋ/ as a phoneme.

    This applies both to borrowings like Tangente and to native words like Wohngemeinschaft
    Not in my dialect. I need a transparent phoneme boundary for “hard n”. Wohngemeinschaft has [ng], Tangente has [ŋg]. (I seem to remember that we already discussed this somewhere in these august halls?)

  226. Can we repeat the trick with /n/?

    “This is not /n/, but /ŋd/ [n]”!

    (or /md/, or otherwise /Nd/ where N is a nasal without a place of articulation).

  227. Tangente has [ŋg]
    At least as long as we’re talking about a tangent. A compound Tangente “seaweed duck” or “duck from the Tang dynasty” would have [ŋ] ([taŋ?εntə]).

  228. David Marjanović says

    That would be an underlying phoneme and therefore cheating.

    That’s why I wrote “otherwise disappeared” and not just “disappeared”. The idea is that every free-standing [ŋ], i.e. not next to a velar plosive, is really /ng/. Assume two steps: first, every /ng/ becomes [ŋg]; second, the [g] is deleted unless it begins a stressed syllable, or, regionally, if certain specific consonants follow.

    In accents that retain phonemic consonant length, “every free-standing [ŋ]” is in fact long, like in Finnish and for the same reason.

    Tangente is a minimal pair for stress, like e.g. umgehen. The ducks have initial stress, so their [g] is deleted, while the tangent has final stress and keeps its [g].

    England has initial stress, and is therefore [ˈʔεŋlant] most widely; southwards, e.g. for me, it’s [ˈ(ʔ)εŋg̊land̥]*, because /l/ blocks the deletion.

    * [ʔ] is only inserted after a pause; the wug kelp duck is [ˈtaŋːεntε] for me

    Wohngemeinschaft

    Ah, now it gets interesting: by the rules I just proposed, the [g] should be deleted, but it isn’t.

    From the speaker’s side, this is trivial: the components of German compound nouns, in this case wohn-gemein-schaft, remain separate phonological words, and the deletion rule doesn’t extend across word boundaries. But of course this doesn’t work from the hearer’s side: rather, the [g] allows the hearer to infer a (phonological) word boundary and thus makes the contrast between [ŋ] and [ŋg] (or, for me, [ŋː] and [ŋg̊]) phonemic.

    The assimilation still happens, though; it doesn’t care about any sort of word boundary, at least where I’m from. Unless I make an effort to speak slowly and clearly, I have [ŋg̊] in Wohngemeinschaft, not [ng̊]. Sebastian Kurz and (Kai) Jan Krainer have been in the Austrian news a lot lately, and they always get [ŋk] on the radio and on TV (public-owned broadcaster in both cases).

    Can we repeat the trick with /n/?

    “This is not /n/, but /ŋd/ [n]”!

    I don’t know, but somebody should try.

    (/n/ and /nː/ do contrast in accents with phonemic vowel length, though.)

    (or /md/, or otherwise /Nd/ where N is a nasal without a place of articulation).

    That does not work, because German does allow /mt/ and (farther down the thread) /md/ in monosyllabic morphemes.

  229. DM and Hans used above “postponed” in the sense of “postposed” (as OPposed to PREposed). It surprised me, I only use the word in the temporal meaning. Now I was reading someone’s grammar and caught myself mentally correcting “postpoSed” to …poNed. (I am not complaining: it was a typo or else “postponed” makes sense here anyway. I am just surprised how it sticked).

  230. It surprised me, I only use the word in the temporal meaning.

    Same here.

  231. Lars Mathiesen says

    The Latin verb is pretty funky, *po-sino/po-situs > pono/positus and hence the different stems in later languages. I don’t think that alternation existed in any other verbs, at least not as reflected in learnèd vocabulary.

    (Spanish still has pongo/puesto, I’m told that ponido is a typical childhood regularization of the latter).

    Anyway, Danish (following German I suppose) only has verbs from -ponere, but of course position.

  232. PlasticPaddy says

    @lars
    The sn/zn > n seems regular (compare e.g., penis). In some verbs like desino or resono maybe there was analogical restoration of the deleted consonant (or the de/re was recognised as a prefix but the po was not)?

  233. David Marjanović says

    My use of postponed and preposed in the same parenthesis looks like I was too tired to pay attention to what my fingers were doing. Many of my typos are the use of common sequences instead of uncommon ones, much like auto-fill-in and autoincorrect actually.

    [Vsn] > [Vzn] > [Vːn] is regular in Latin.

  234. Yes, I thought the same. I am just surprised how easily it sticked.

    Many complaints at the falling literacy (spreading illiteracy) in Russia are caused by misspelling of omophonic sequences (the English version is you’re/your). I do it myself when I type fast, in words whose spelling I know well: apparently, at a certain speed my fingers receive phonological information about a word’s elements rather than a word as a whole. I never made these typos when I was writing.

    Unstressed a/o (кАрова) and (rarely in my case. I only did that after many years of typing fast) infinitive and third person endings (она мне нравитЬся). This soft sign is the most popular mistake.

    So the people who complain were saying “what a shame! They don’t even know how to write [….]” and I was saying “those are typos, not errors” and they were saying “no, those are errors, not typos! What a shame!”

  235. People love being indignant about other people’s mistakes.

  236. But the “typos” theory is interesting: it tells us something about how speech is processed in our brains. I only can explain my typos with that my “phonological processor” that usually gives commands to my tongue starts doing that to my fingers as well when I increace speed. I can’t explain why it does not happen when I write.
    In my case it is not a theory, it is a fact (compare to slow typing, compare to writing with a pen, and then I just never made orphographical mistakes in school).

    The “illiteracy” theory is less interesting. Yes, now we have the Internet and people who did not read and write often in Soviet times began writing on forums.

  237. David Marjanović says

    I can’t write by hand anywhere near as fast as I can type, so that would explain why I, too, make phonological mistakes of that sort when I type. However, I seem to catch all of them because I look at the screen when I type. I don’t catch all of the others in real time – similar to how autoincorrect can’t tell if you meant form or from, I guess.

  238. Lars Mathiesen says

    I don’t doubt that there are other words where that sound law applied, but the present stem of sinō/sivī/situs/sinere has the nasal infix so the change only happens there, unlike 1st conjugation sonō/sonuī/sonitum/sonāre — that’s what makes the alternation so funky. (If resonō > rēnō had stuck, I assume it would have been rēnuī/rēnitus/rēnāre without any funk).

    There may have been other verbs in Classical Latin where the same thing had happened, but I think pōnō/positus is the only one that survived to confuse us. Analogy must have worked on desinence at some point, like PP says (and anyway there are no words around from the pptc stem to compare to, I don’t think).

  239. The four periods of my life are:

    – reading and writing a lot, never making mistakes (school)
    – reading mostly in English, typing carelessly on Russian forums. Never caring about punctuation, as a conscious decision.
    – reading mostly in English, typing carelessly on Russian forums, making many more mistakes and forgetting Russian punctuation, because I am used to typing carelessly.
    – reading and speaking a mixture of languages, God knows what my Russian looks like now. I think I begin to forget some spellings.

    Yes, “writing by hand is slower” is a possible explanation (there are other differences like: it is connected). It is somewhat hard to believe, because at the stage 1 my writing was just perfect while at the stage 2 I still had not learned to type as fast as today. But maybe it is true.

    A weird thing happened to девчонка. I always wrote it correctly and the incorrect spelling looked just odd. Then some guy on a forum began complaining at people who write девчЁнка. This spelling irritated him. At first I was able to understand him (no, it did not irritate me, but it looked odd). But then my input (with respect to this particular word) wholly consisted of his complaints. I began seeing both forms equally often and soon I found that I can’t remember which one is what irritates him and which one is what is “correct”.

  240. Lars Mathiesen says

    My fingers like to type “chack” instead of “check”. It’s a little interesting that it’s a vowel for vowel substitution, but I can’t find any phonological excuse. (I still know how it’s supposed to feel when I type “check,” so I usually know I typed it wrong without looking).

    Maybe it’s because in my personal perversion of touch typing I use the same finger for “e” and “c” but not for “a” — so the sequence “ac” is somehow easier than “ec”.

  241. I am not sure if I have a good theory of distinction between “errors” and “typos” thoguh.

    Imagine, you asked a person (a 5 years old girl or an adult who never read books and writes with errors) about the spelling of a certain form, and this person gave a correct answer. Apparently the proper spelling is available to her “logical” mind, but not to her “authomatical/subconscious” mind.

    Compared to her I reproduce it authomatically, this is why I prefer the word “typo”. But my “logical” mind thinks nothing about this spelling, I just write it properly because I see it in books. In that I am similar to other people who read book in childhood. She in turn is similar to younger kids or “uneducated” people. We can use my instinctive knowlege as a class/age marker.
    But then:
    – can we say that she is making “errors” rather than “typos”? She knows the proper spelling.
    – can we say that I am more “literate”? She can be better versed in the rules that school teachers teach, especially if she was getting bad grades.

  242. if
    – “a literate person making typos” is someone who feels that a wrong spelling is “odd” and is the correct spelling is “normal” and, accordingly, to make mistakes less often needs to watch what she is typing
    and
    – “an illiterate person making errors” is someone who does not feel that, and, accrodingly, to make mistakes less often tryed to change her subconscious representation with will power (does not work) or mnemonics or what not (other methods that do not work…)

    then the terrible story of девчٍٍVнка is a case of acquired illiteracy (as opposed to literacy not acquired).

    (sorry, I am just thining aloud).

  243. devčonka:

    We have devčata “girls”. It presupposes neuter singular *devča (singular oblique *devčati). Cf. Czech n. děvče.

    Likely devčonka was formed from it by analogy with words like PS *žerbę pl *žerbęta “foal”, Russian dated neuter жеребя pl. жеребята with neuter diminutive *-en(t). The neuter singular is not used (except дитя “child”), instead we formed masculine singular жеребёнок “foal”, while retaining the plural жеребята. -ёнок is now understood as a suffix for animal kids.

    devčonka now has two plurals:

    pl devčata > sg. devčonka > pl. devčonki

  244. or else -ёнок forms are old.
    They have this -n-, which is hidden VERY well in modern Russian -я Proto-Slavic nasal ę… In that case they just lost their plural and united with neuter forms that lost singular.

  245. David Marjanović says

    How does ребёнок pl. ребята fit into this?

    Never caring about punctuation, as a conscious decision.

    I’m not capable of that.

    My fingers like to type “chack” instead of “check”. It’s a little interesting that it’s a vowel for vowel substitution, but I can’t find any phonological excuse.

    I sometimes exchange(d?) a and e in typing, too, likewise not for any phonological reason (in German especially). Maybe it’s something neurological about the left hand.

  246. The idea is that every free-standing [ŋ], i.e. not next to a velar plosive, is really /ng/.
    Well, that’s its historical origin and also what the spelling indicates, and it’s also fine as a morphonological rule, but IMO it’s wrong on the level of phonological analysis, because it confuses levels of analysis.
    I am aware that two phones that are in complementary distribution with each other are normally seen as allophones one phoneme, but [ŋ] is in complementary distribution with [ŋg] only in some environments; in others it’s in complementary distribution with [n], and I don’t share analyses where one phone is an allophone of two different phonemes.

  247. Like David I read as I type and usually live-correct. But homophones often get me. I left it’s for its here the other day and then slit my wrists when I later saw it. “They’re going to think I’m the kind of person…”

    I think it’s because though I type faster than I handwrite or speak, I’m still thinking faster and somehow processing the words individually from residual short-term memory which is probably audio after a fashion.

    Also, its much more likely to happen when phone-keying, maybe because I’m much slower, so the word is that much more distant from it’s thought, and the balance between homophony and meaning has tilted even further towards sound.

    (Its/it’s confusion above is intentional.)

  248. David Eddyshaw says

    Those errors in my own comments which become apparent to me five seconds after the editing period ends, seem (other than simple incoherence) to consist predominantly of complete omission of function words. I expect this proves something about Universal Grammar. Or not. As you may think.

  249. I left it’s for its here the other day and then slit my wrists when I later saw it.

    For what it’s worth, I do that all the time (make the error, that is — I usually notice and correct it). I suspect that’s the most common typo in English, for fairly obvious reasons.

  250. “I expect this proves something about Universal Grammar.” I am most amused.

    > Ryan: I think it’s because though I type faster than I handwrite or speak, I’m still thinking faster and somehow processing the words individually from residual short-term memory which is probably audio after a fashion.

    It’s the same with me, I think, although it’s not auraul but grphical.

  251. I suffered an injury on my left wrist recently, and it is interfering with my muscle memory for typing. It does not seem like it will recover.

  252. My most common typos are omitting letters by not pressing the keys hard enough, followed by homophones. That aside from incomplete editing (e.g. moving a word around but not deleting its original occurrence) and bad grammar.

  253. The pinkie finger of my left hand goes too far to the left, and the ring finger too far to the right. The pinkie finger, that I usually use to press the function keys, is now practically useless at the best of times, and getting in the way at the best of times. Also the ring finger is also not very useful. I can’t make the Vulcan salute any more with my left hand 🙂 The middle fingers is still moderately useful, but with a lot less strength than before. An year and a half now.

  254. David Marjanović says

    I don’t share analyses where one phone is an allophone of two different phonemes

    I don’t have a problem with analyses where a single sound is how a sequence of two (or perhaps more) phonemes is pronounced (under all or some circumstances); off the top of my head, [ʃː] as /st͡ʃ/ in conservative Standard Italian seems to work. Where the analysis of [ŋ(ː)] as not phonemic breaks down is the “postlexical level”, the hearer’s point of view.

    Lots of such analyses break down at that level, e.g. [tʰ] and [t̚ˀ] as allophones in English as we discussed recently.

  255. How does ребёнок pl. ребята fit into this?

    It is the same. orbę.

    In Russian in plural робята 15th century, singular from 17th century with -e- and -o-.
    In West Slavic in Polish in a 15th century Bible robyenci, robyonkowye, robyonkem and robota (understood as misspelling of robyøta < robię (whatever ø means)).
    I do not know what are early Czech attestations:(

    The -o- form looks exactly as the stem of robъ / roba / rabъ / raba “slave” with the neuter *-en(t) suffix.
    Are there any Germanic parallels with X “worker, slave” and diminutive of X “child”?

  256. dravsi: in Bulgarian rabotnik etc. just means worker. Why do you think it has shades of enforced labour in northwest and northeast Slavic?

  257. @V, no, it does not.
    A “slave” is раб, as in Bulgarian. Работник is worker in a more generalized sense (e.g. научный работник), while “worker” as a member of working class, a factory worker, is рабочий, a substantivated adjective. It happens to professional jargons: e.g. circus artists identify as цирковые (a substantivated adjective), while normal Russian is циркач.

  258. I just mean, there is a pair
    slon “elephant”, slonyata “young elephants”. If you like, you can form Klingonyata for “Klingon children”.

    And a pair rob “a variant of modern rab, slave”, robyata “a variant of modern rebyata, children” looks similar. But I do not mean that robyata/rebyata necessarily was formed based on the word робъ in the sense “slave”.

  259. Child/slave parallels are not unknown with other words, e.g. Czech otrok (slave) versus Slovene otrok (child). Could lead to misunderstandings…

  260. also
    https://en.wikipedia.org/wiki/Kholop
    https://en.wiktionary.org/wiki/Reconstruction:Proto-Slavic/paorbъkъ

    All in the same region. I did not think about otrok !

    I asked about Germanic because (1) it seem to happen in around Baltic (2) Germanic has a suffix similar to *-en(t) neuter diminutive (and German “girl” is neuter).

  261. Child/slave parallels are not unknown with other words

    In North Africa one can find several instances of child/bug parallels, such as Korandje lhaybuš “children” from a dialectal Arabic word meaning “lice”. And I’ve heard namusa “mosquito” used as an endearment to toddlers…

  262. PlasticPaddy says

    @lameen
    Mosquito is good, but depending on the pitch of the child, sometimes police siren might be a better match.

  263. PlasticPaddy says

    @drasvi
    German and English seem to use plant growth metaphors, i.e., offspring (i think offshoot can also be used metaphorically), Sprössling, Nachwuchs.

  264. Reminded me to look up the etymology of “little buggers”.

    PS I looked.
    It is from “Bulgarians”:-/

  265. David Eddyshaw says

    Plant.

  266. Weeds:-E
    Mala herba cito crescit.

  267. A!

    Someone said that a common joke is that to speak Yiddish you need to add “a” to every word (someone is a a Russian speaking Jew addressing Russian-speaking Jews).
    The idea is that this “a” is a salient feature, appearing in contexts where English would not use it.

    I never heard the joke. But I certainly heard the “a”. So we have a massively indefinite language (as opposed to l’-languages.)

  268. David Eddyshaw says

    I can’t think of any language which has an indefinite article but no definite article, though the reverse is common enough, of course. Did Greenberg get around to that one?

    (I’m not altogether sure how you could be certain that something actually was an indefinite article if there was no definite article to contrast it with, though, so this may be an illusory universal, even if no clearcut counterexamples emerge …)

  269. David Marjanović says

    “One” + classifier is used in Mandarin quite a bit more often than necessary. Pinyin seizes on that and spells this as one word for the most common classifier. At the same time, there is no hint of a definite article.

  270. David Eddyshaw says

    Cheating slightly, I suppose Proto-Western-Oti-Volta might have been such a language: whereas the various definite articles of the modern WOV languages are of different origins, the specific-yet-indefinite dependent pronoun (sɔ’/si’a/sieba in Kusaal) certainly goes back to Proto-WOV. The meaning is not just “indefinite”, though: it’s “a certain”, or quite often “another” (just like Hausa wani/wata/wasu.)

  271. Two of my everyday languages, Kurmanji and Turkish… This map has the value “No definite, but indefinite article” as a white diamond:

    https://wals.info/feature/37A#2/25.5/149.1

    Literary New Persian as well. There is a description of the situation in modern colloquial Iranian Persian here (link to pdf):

    https://langsci-press.org/catalog/view/283/2676/2003-1

  272. David Eddyshaw says

    Yup. I was comprehensively wrong. Greenberg was too fly to fall for that one …

    I suppose I could try to rescue the hypothesis by wittering about definite vs indefinite marking (not necessarily by articles), which has the advantage of being so open to interpretation that it would allow you to explain away any inconvenient facts at all. Perhaps not …

  273. I don’t have a problem with analyses where a single sound is how a sequence of two (or perhaps more) phonemes is pronounced (under all or some circumstances)
    That’s not what I meant. What I meant is that if you want to deny phoneme status to [ŋ], then in some positions you have to describe [ŋ] as allophone of /n/, and in others (where it contrasts with /n/) as allophone of the phoneme combination /n/ plus /g/.

  274. rescue the hypothesis by wittering about definite vs indefinite marking (not necessarily by articles), which has the advantage of being so open to interpretation that it would allow you to explain away any inconvenient facts at all

    I was surprised to see Japanese categorized as having an indefinite article in the map on WALS that I linked to… I have never really used the site, so I investigated. According to the mouse-over reference, this categorization is based on John Hinds 1986 Japanese: Descriptive Grammar, p. 81, which is this:

    1.2.5.2.4. ARTICLE
    There is no article, although a three-way demonstrative system exists to show definiteness. Indefiniteness may be shown by aru “a certain”. These all precede the nominal: sono hito “that person”, aru hito “a certain person”.

    So that’s what can get categorized as an indefinite article in WALS…

  275. Then Russian некий is an indefinite article.

  276. David Marjanović says

    Are there any Germanic parallels with X “worker, slave” and diminutive of X “child”?

    Sort of. Mädchen “girl” is from Mägdchen (not sure if that’s attested), the diminutive of Magd f., which is the cognate of maid and has had the same range of meanings: “young woman; unmarried woman (e.g. St. Mary); handmaid, milk- or other maid on a farm” (only the latter is current).

    That’s not what I meant. What I meant is that if you want to deny phoneme status to [ŋ], then in some positions you have to describe [ŋ] as allophone of /n/, and in others (where it contrasts with /n/) as allophone of the phoneme combination /n/ plus /g/.

    Ah yes. In my accent, these two differ by length, so that [ŋ] would be /n/ and [ŋː] would be /ng/, but of course that trick doesn’t work for yours!

    (Phonemicization of /ŋ/ caused by loss of consonant length. Who’d’a thunk it.)

    …though even so, “[ŋ] is /n/ if it’s next to a velar plosive, /ng/ if it’s not” would still work. Except across word boundaries of course.

  277. David Marjanović says

    I missed an apostrophe in who’d’a’. 🙂

  278. David Eddyshaw says

    So that’s what can get categorized as an indefinite article in WALS…

    I find WALS unreliable uncomfortably often whenever I’m in a position to check.

    In this particular case, the “affix” category looks more than a bit arbitrary. Danish has a “definite affix”, I see; so has Icelandic, where this “affix” actually declines.

    Mooré is described as having a “definite affix”, but this “affix ” is actually simply a specialised use of the word “this” (which can head a NP all by itself) with an entirely regular sandhi loss of the /w/. The same thing happens with the cognate word nwa in Kusaal whenever it is used as a postposed dependent, though Kusaal has actually adopted the other deictic, la “that” as its article. In Toende Kusaal, the l- of la, when used as an article, is lost after any word-final consonant, but not after vowels. When the l is lost, the article is written as an “affixed” -a.

    In both Kusaal dialects, la also remains “that” (as opposed to “this”) in contexts where the head is already definite (for example, after a definite possessor, or along with a demonstrative, most of which do not themselves distinguish near/far.) In other words, only syntactic context tells you whether “the” is distinct from “that/this”; it’s not a property of the article-word in itself, so there is no real answer as to whether the article is “distinct from a demonstrative.”

    In both Mooré and Kusaal, the “affixed” article is attached to the last word of the definite NP, not to the head noun.

    In a nutshell, all these “affixes” are without doubt really words – unless you deny that status to English and French articles as well.

    I think the problem is partly that WALS seems to rely heavily on secondary sources, which often incorporate questionable predigested analyses (sometimes shoehorned into predetermined categories whether they really fit or not) of primary sources which are not always all that reliable in the first place.

  279. …though even so, “[ŋ] is /n/ if it’s next to a velar plosive, /ng/ if it’s not” would still work.
    I’m repeating myself, that’s the kind of rule I don’t accept. For me, a phone that is an allophone of two different phonemes in different positions (so I accept neutralizations) is a phoneme. On top of this, I prefer not to mix historical explanations / reading rules with phonology. We know that [ŋ] goes back to [ng] historically and that “ng” is pronounced [ŋ] in certain positions, but on the synchronic level of phonological description there is only /ŋ/, which is a phoneme. If we proceed otherwise, we might as well postulate the “underlying” i’s causing umlaut and similar excesses perpetrated by generative phonology.
    Except across word boundaries of course
    So, while for you that’s the only / decisive argument for a phoneme /ŋ/, for me it’s just an additional argument.

  280. David Marjanović says

    If we proceed otherwise, we might as well postulate the “underlying” i’s causing umlaut and similar excesses perpetrated by generative phonology.

    Umlaut is so obviously morphological and lexical – Kraft, pl. Kräfte, but Haft, pl. Hafte; hart, comparative härter, but zart, comparative zarter (hart & zart don’t rhyme for everyone, though); dunkel, comparative mostly dunkler, but dünkler in Vienna – that any such postulate would require a huge amount of epicycles at least on par with Chomsky & Halle’s mainstream English /x/ only ever surfaces as a switch from one vowel phoneme to another. Postulating that free-standing [ŋ], let alone the southern [ŋː], is /ng/ is much less unparsimonious.

  281. David Eddyshaw says

    On plants, slaves, bugs, mosquitoes and children: Kusaal has zɔrʋg “piece, crumb” beside zɔrig “piece, crumb, child” (same stem, different noun classes.) The “child” sense isn’t in the dictionaries but was given to me by a reliable informant; I suspect it’s more or less like “sprog” in general tone.

    [The first word exactly matches Mooré zorgo “a species of plant” from an etymological standpoint, both segmentally and tonally, but that doesn’t necessarily mean anything, because (non-Boulba) Western Oti-Volta /z/ has two distinct origins in Proto-Oti-Volta; similarly with WOV /s/: delightful as it would be if siig “spirit” were really “the same word” as the exactly homophonous siig “African birch tree”, sadly, they definitely have quite different origins.]

    Nobody in Oti-Voltaland seems to use the “child” word (*bi-) for “slave”, or vice versa. The same “slave” word turns up everywhere in the group (Kusaal has yammʋg) but it’s structurally peculiar for an inherited Oti-Volta word and the sound correspondences often don’t seem to work right. I suspect it’s got loaned from one Oti-Volta language to another (which would, alas, make some sense with a word meaning “slave.”)

  282. Is Voltaland in any way connected to “Upper Volta”?

  283. Any sense for a word meaning slave is alas.

  284. David Eddyshaw says

    Is Voltaland in any way connected to “Upper Volta”?

    Yes.

  285. The all are rivers.

    Congo, Ubangi, Niger, Volta, Oti, Benue/Tchadda, Senegal, Gambia…
    Cameroon (Rio dos Camarões)
    But not Angola. Not Guinea. Not Benin/Dahomey

  286. David Eddyshaw says

    As drasvi says …

    The names for Niger-Congo subgroups are mostly consciously taken from rivers (and the Atlantic.)
    The supposed subgroup called “Gur” in English is “Voltaïque” in French, which is frankly a much better name; I think it was avoided in English because of potential confusion with “Haute Volta”/”Upper Volta” [also named for the river(s), of course] which is no longer an issue, thanks to Thomas Sankara. And anyway, that doesn’t stop Chadicists …

    The Oti is a tributary of the Volta:

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

    I think “Gur” is actually not a real subgroup, but Oti-Volta certainly is. There’s been a push, led, I think, by Adams Bodomo, to use “Mabia” instead, which is reflected in WP to some extent, but his own classification is very impressionistic, discarding much of what is actually known about the interrelationships of the languages, and largely ignores work published in French (which is actually most of the relevant stuff.) His acolytes also use the term inconsistently among themselves, sometimes meaning just “Western Oti-Volta” (which does indeed need a better name) but usually including Buli and often also the Gurma languages, and even Kasem, which is not even an Oti-Volta language; sometimes it’s used as a substitute for all of “Gur.”

    I also don’t like the term itself: “mabia” is Dagaare for “full sibling” (Bodomo is Dagaba), but even other WOV languages have different forms (Kusaal mabiig) and outside WOV the equivalents are quite different. As a “shibboleth” word it doesn’t actually work. I do like the idea of using an indigenous name in principle, though, if someone could come up with a suitable one not too tied to a particular pet language.

  287. David Eddyshaw says

    Sadly, my suggestion of “Tarim-Liffey” to replace the awkward and inaccurate term “Indo-European” has not yet been widely adopted.

  288. What about Kra-Dai, rather than Tai-Kadai? Kra-Dai seem to have mended things, seeing as Kra languages seems to have been the Urheimat. I have a Hmong acquaintance who prefers that.

  289. David Marjanović says

    The Hmong-Mien languages are only related to Kra-Dai if you accept the Austric macro-proposal, though. Kra-Dai is in any case closer to Austronesian.

  290. Baxter and Sagart:

    We adopt the term “Kra-Dai” proposed by Ostapirat (2000) in place of the traditional “Tai-Kadai,” since to Thai speakers, “Tai-Kadai” evidently sounds unintentionally funny, meaning something like “Tai, or whatever” (Montatip Krishnamra, p.c.)

  291. Kra-Dai is in any case closer to Austronesian.

    Says Sagart, but he’s not convinced enough people. Blust (p. 710): “Few linguists viewing these comparisons would feel comfortable attributing them to chance. Whether the historical connection is due to contact or divergent descent, however, remains a point of contention.”

  292. @David Marjanović : everything else considered, how can Kra-Dai be related to Austronesian?

  293. Oceanic should be renamed Muvian.

  294. @Y: I don’t think Oceanic is a thing.

  295. Tell that to the whales.

  296. David Eddyshaw says
  297. David Marjanović says

    It’s not just Sagart; Sagart is the one who puts Kra-Dai inside Austronesian, right next to Malayo-Polynesian.

    If I find the YouTube video(s?) again, I’ll post them…

  298. جنس

    Arabic
    Etymology 1
    Probably from Classical Syriac ܓܸܢܣܵܐ‎ (gensā), which is in turn probably from Ancient Greek γένος (génos), from Proto-Indo-European *ǵénh₁os. Alternatively, but much less likely, from Latin genus from the same Indo-European etymon. For the word’s meanings, compare with German Geschlecht (“race; genus; maleness or femaleness; grammatical gender”).

    جِنْس • (jins) m (plural أَجْنَاس‎ (ʾajnās))

    1. (countable) a kind; a variety; a breed

    ‏اَلْجَزَاءُ مِنْ جِنْسِ الْعَمَل‎

    al-jazāʾu min jinsi l-ʿamal
    The punishment fits the crime.

    Synonyms: صَنْف‎ (ṣanf), نَوْع‎ (nawʿ), نَمَط‎ (namaṭ), طِرَاز‎ (ṭirāz), فِئَة‎ (fiʾa), فَصِيلَة‎ (faṣīla), سُلَالَة‎ (sulāla)

    2. (countable, philosophy) a category to which one or more things belong

    Synonyms: مَقُولَة‎ (maqūla), مَحْمُولَة‎ (maḥmūla)

    1) (see usage notes) the quality of maleness or femaleness
    2) (grammar) grammatical gender
    3) (of animals or plants) a stock; a breed

    Hyponym: (of humans, modern) عِرْق‎ (ʿirq, “race”)

    4) (taxonomy) genus

    Coordinate terms: نَوْع‎ (nawʿ, “species”), مُسْتَنْبَت‎ (mustanbat, “cultivar”), فَصِيلَة‎ (faṣīla, “family”), رُتْبَة‎ (rutba, “order”), مَمْلَكَة‎ (mamlaka, “kingdom”)

    3. (uncountable, modern, sometimes vulgar) sexual intercourse

    Synonyms: مُضَاجَعَة‎ (muḍājaʿa, “lying together; lovemaking”), جِمَاع‎ (jimāʿ, “copulation”), (archaic) نِكَاح‎ (nikāḥ), (now vulgar) نَيْك‎ (nayk)

    4. value, worth, distinction which reflects in a different price, richness
    5. good or item of value of a furnishing, a valuable, commodity
    _________
    I wonder if there are etymological sources with fewers probablys.

  299. >I don’t think Oceanic is a thing.

    Of course it is. The real question is whether the Oceanic languages are currently more closely allied with the East Asiatic languages or the Eurasiatic languages.

  300. David Eddyshaw says

    They’ve always been more closely allied with the Eurasiatic languages.

    [Do not be taken in by the deceptions of Joseph Greenstein!]

  301. David Marjanović says

    German Geschlecht (“race; genus; maleness or femaleness; grammatical gender”).

    Uh… “noble/ancient family” (whence Polish szlachta “gentry ~ nobility”, i.e. people who belong to a named Geschlecht); Menschengeschlecht “the human race” (poetic); “sex ~ gender” (in biology, on forms, in grammar). I think that’s it. “Genus”, in taxonomy, is Gattung.

  302. Geschlecht

    I’ve just realized that Swedish slag and släkt are distantly related.

    slag
    Etymology
    From Old Norse slag, from Proto-Germanic *slagiz (“hit, blow”).
    […]
    7. A kind; sort.

    En fågel av ett ovanligt slag ― A bird of an unusual kind
    Synonym: sort

    släkt c
    Etymology
    From Old Swedish slækt, borrowed from Middle Low German slecht (“family, lineage”), from Proto-Germanic *slahtō, from the verb *slahaną.

    1. kinship, relatives

    hela tjocka släkten

    all my relatives

  303. January First-of-May says

    What about Kra-Dai, rather than Tai-Kadai?

    dodge the Tai/Thai problem entirely by saying that the Siamese language is in the Shan-Zhuang branch of Austrepeiran

    (not actually sure which, if any, of those corresponds to Kra-Dai)

  304. David Marjanović says

    I’ve just realized that Swedish slag and släkt are distantly related.

    …and the same in German, which I hadn’t noticed either!

  305. PlasticPaddy says

    There is an apparent cognate “sliocht” in Irish:
    1. Mark, trace, track.
    2. Offspring, progeny; line, posterity
    3. Passage, tract; extract.
    Apparently 1 is the “original” sense, although 2 at least goes back to Old Irish
    Matasović has for Proto-Celtic
    *slig-o- ‘strike, hew’ [Vb]

    GOID: OIr. sligid, -slig; sleiss, -sle [Subj.]; silis, -sil [Fut.]; selaig [Pret.];
    slechtae, -slecht [Pret. Pass.]
    PIE: *sleyg- ‘smear, creep’
    Matasović says:
    ETYM: The meanings ‘smear’, ‘sneak’ and ‘hit’ are not easily reconciled,
    but cf. Germ. Streich ‘blow’ and streichen ‘smear’.

  306. OIr. sligid

    Well known to any beginning student of the language!

  307. Could Irish sliocht instead be related to English slick and German schleichen/OHG slihhan?

    Maybe they’re all tangled up together, along with sleigh and slide, through some meaning of line, path.

  308. John Cowan says

    wrt the ‘Hearing Faculty’ all humans have (or learn) an amazing ability to tell the direction a sound is coming from, merely from the timing difference in when it reaches one ear vs the other.

    In that case my father and my wife do not count as human (and what does that make me?)

    The Tractatus is indeed as beautiful as it is wrong.

    It’s poetry, as DM pointed out (“Die Welt ist alles, was der Fall ist”): how can it be said to be wrong? The poet never affirmeth.

    although the actual people of Ugarit did not regard themselves themselves as Canaanites

    By the same token, the actual people of the United States do not regard themselves as English (except those who actually are English, of course).

  309. David Marjanović says

    The poet never affirmeth.

    Erasmus Darwin did, so did Lucretius, and so did Wittgenstein as far as I understand.

  310. David Eddyshaw says

    so did Wittgenstein

    Yes indeed, at least in that work; although the twist (SPOILER!) is that he unaffirms at the end.

    I think I’ve cited Lucretius myself previously in this connexion.

  311. Not to mention Walt Whitman, who was constantly affirming.

  312. John Cowan says

    Let’s have a bit of context from the Defense of Poesy, then, where Sidney is concerned to defend “poor poets” from the charge of lying:

    To the second [charge], therefore, that they [poets] should be the principal liars, I answer paradoxically, but truly, I think truly, that of all writers under the sun the poet is the least liar; and though he would, as a poet can scarcely be a liar. The astronomer, with his cousin the geometrician, can hardly escape when they take upon them to measure the height of the stars [this is 1579]. How often, think you, do the physicians lie, when they aver things good for sicknesses, which afterwards send Charon a great number of souls drowned in a potion before they come to his ferry? And no less of the rest which take upon them to affirm.

    Now for the poet, he nothing affirmeth, and therefore never lieth. For, as I take it, to lie is to affirm that to be true which is false; so as the other artists, and especially the historian, affirming many things, can, in the cloudy knowledge of mankind, hardly escape from many lies. But the poet, as I said before, never affirmeth. The poet never maketh any circles about your imagination, to conjure you to believe for true what he writeth. He citeth not authorities of other histories, but even for his entry calleth the sweet Muses to inspire into him a good invention; in troth, not laboring to tell you what is or is not, but what should or should not be.

    And therefore though he recount things not true, yet because he telleth them not for true he lieth not; without we will say that Nathan lied in his speech, before alleged, to David; which, as a wicked man durst scarce say, so think I none so simple would say that Æsop lied in the tales of his beasts; for who thinketh that Æsop wrote it for actually true, were well worthy to have his name chronicled among the beasts he writeth of. What child is there that, coming to a play, and seeing Thebes written in great letters upon an old door, doth believe that it is Thebes? If then a man can arrive at that child’s-age, to know that the poet’s persons and doings are but pictures what should be, and not stories what have been, they will never give the lie to things not affirmatively but allegorically and figuratively written. And therefore, as in history looking for truth, they may go away full-fraught with falsehood, so in poesy looking but for fiction, they shall use the narration but as an imaginative ground—plot of a profitable invention.

    But hereto is replied that the poets give names to men they write of, which argueth a conceit of an actual truth, and so, not being true, proveth a falsehood. And doth the lawyer lie then, when, under the names of John of the Stile, and John of the Nokes [variants of John Doe and Richard Roe], he putteth his case? But that is easily answered: their naming of men is but to make their picture the more lively, and not to build any history. Painting men, they cannot leave men nameless. We see we cannot play at chess but that we must give names to our chess-men; and yet, me thinks, he were a very partial champion of truth that would say we lied for giving a piece of wood the reverend title of a bishop. The poet nameth Cyrus and Æneas no other way than to show what men of their fames, fortunes, and estates should do.

    And so. I say, if Lucretius his works have any merit today, it is not and cannot be because of his Physics.

Speak Your Mind

*