I’ve started reading A Social History of Hebrew: Its Origins Through the Rabbinic Period, by William M. Schniedewind (thanks, bulbul!), and I’m already learning new things. Here’s a passage about the name of the language:

Finally, it should be noted that the term Hebrew (עברית /ʿiḇrîṯ/) itself is not a biblical term for the Hebrew language. Although we now call the language of ancient Israel and Judah “Hebrew,” this term first appears for the language in the Mishnah (m. Gittin 9:6, 8; m. Yadayim 4:5), which was edited around 230 C.E. That is, the metalinguistic term Hebrew emerged precisely when the speech community in Palestine was disappearing. Furthermore, it appeared when the traditional linguistic identification of people, language, and land had disappeared with the changed sociopolitical situation of the Jewish community [after the Bar Kokhba revolt and the expulsion of the Jews from Judah/Judea]. This new term also formally recognized the difference between Hebrew and Aramaic as two Jewish languages, in contrast to the New Testament’s use of the Greek word hebraisti (Ἑβραϊστι) to refer to the vernacular Jewish language, without making clear whether this term refers to Hebrew or Aramaic. The term for the Hebrew language that would have been used by the prophet Isaiah in 700 B.C.E or the priest Ezra in 400 B.C.E is Yehudite (יהודית /yehûḏîṯ/), which is derived from the word for the territory of Judah (יהודה /yehûḏâ/) and is also generally used for the ethnic group Judean/Jew (יהודי /yehûḏî/) in postexilic biblical literature, Qumran literature, and rabbinic literature. It is quite typical of languages to be called after territories and ethnicities: thus, German and Germany, English and England, or Chinese and China. This fact actually highlights changes in the Hebrew language as it related to Jewish identity in different periods. The Jews/Judeans who lived in Judah/Judea always spoke the Judean/Jewish language. It is only when the Jews were expelled from Judea that the Judean language ceased to be a living vernacular. In fact, it is only at this time that the Jewish language became “Hebrew.” When the terminology for the language of the Jews is separated from that of the territory, it marks a profound shift in the history of the language itself.

Schniedewind won my heart earlier in the chapter when he wrote: “Here, it is important to recognize that all language classification is shaped by linguistic ideologies. For example, the description of Chinese as one language with several dialects, and of Norwegian, Danish, and Swedish as three languages is more a reflex of nationalism and borders than the conclusion of descriptive linguistics.” I’m really looking forward to the rest of the book!


  1. Huh, this is eerie. The main example I know of the language being called “Yehudit” is in 2 Kings, 18 verse 26 (set during the Assyrian siege of Jerusalem, ~720 BC). The translation I found, though, goes thusly:
    Then Eliakim son of Hilkiah, and Shebna and Joah said to the field commander, “Please speak to your servants in Aramaic, since we understand it. Don’t speak to us in Hebrew in the hearing of the people on the wall.”

  2. J.W. Brewer says

    The relevant section of the LXX has “λάλησον δὴ πρὸς τοὺς παῖδάς σου Συριστί, ὅτι ἀκούομεν ἡμεῖς, καὶ οὐ λαλήσεις μεθ᾿ ἡμῶν ᾿Ιουδαϊστί” etc. So it was not impossible to distinguish Aramaic from Hebrew in Greek, even if the NT authors did not do so. Yuval, fwiw, the translation you are referencing is the NIV, which is rather free (some might say so free as to fall on the other side of the admittedly blurry line between translation and paraphrase), so you might want to check against some other recentish versions, which use locutions like “the language of Judah” (NRSV and ESV) or “Judean” (NASB, and also how NETS renders the LXX).

  3. In two or three posts, because that’s the way WordPress wants it:

    The following, with minor changes, is from Section 1, ‘Origin of the name Hebrew’, in the Hebrew-language Wiki entry on that language.

    “The name עֵבֶר Ever [often spelled Eber but that’s not the way I learned how to pronounce it way back — PO) appears in the Bible as the name of an ancestor of our forefather Abraham. The term עברי Ivri is mentioned many times in the Bible as a name for the Children of Israel, yet the name of the language of the Hebrews is called in the Bible יהודית Yehudith (footnote here leads to several examples, the earliest of which is 2 Kings 18:26: ‘Speak Aramaic to your servants, for we understand it; do not speak Yehudith among us . . . and he proclaimed in a loud voice YEHUDITH.’ The KJV has “Syrian language” rather than Aramaic — PO) and also in the Book of Isaiah, 19:18, as ‘the language of Canaan’. This name for the language was probably used to differentiate the dialect of the (northern) Kingdom of Israel from that of the Kingdom of Judah, or that of the area around Jerusalem.

  4. “The oldest source using the word Hebrew for the language is a Greek source from the 2nd century BC. The source is the grandson of בן סירא Ben Sirach, who writes in a foreword to a translation into Greek of his grandfather’s composition: “Wherefore let me entreat you to read it with favor and attention, and to pardon us, wherein we may seem to come short of some words, which we have labored to interpret. For the same things uttered in הֵבְּרַאִיסְטִי ἑβραϊστὶ Hebrew, and translated into another tongue, have not the same force in them.

    “In the language of our sages (Talmud scholars — PO), spoken from the beginning of the 1st century AD, the accepted term for the language of the Bible and of the spoken language was לשון הקודש ‘holy tongue,’ and the term ‘Hebrew’ was used to designate the old Hebrew script, and to designate the language of ‘the children across [ עבר ever ] the river,’ although there are instances where the term ‘Hebrew’ also serves to designate the language of the Bible.

  5. עבר avar is an everyday term in Modern Hebrew that means ‘to cross, go over’ and similar. It also designates the grammatical past tense, and with a twist, to mean ‘to forward,’ as in ‘I forwarded the email to several friends.’ It’s also used to mean a reworking of a text into another language rather than a word-by-word translation.

    I’m not sure when the term יהודים Jews came to be. Throughout the Bible and Talmud, and seen today in prayers and a host of everyday sayings, the term is עם ישראל People of Israel or Nation of Israel.

  6. The Hebrew Wikipedia article has a good section on the history of the term. Summing it:
    — עברי ‘Hebrew’ comes from עֵבֶר ‘Eber’, an ancestor of Abraham. I don’t know of any suggestion that he name is related to the Hebrew root עבר mentioned by Paul Ogden. Rashi interprets some Talmudic uses of the word עברית as ‘the language of the people across the river’, deriving from that root, but I don’t see it.
    — The first mention of ‘Hebrew’ as a name for the language is in the introduction to the Greek translation of the Book of Ben Sira (Wisdom of Sirach). The translation and the introduction, dated to the 2nd century BCE, are by Ben Sira’s grandson, who writes, “Wherefore let me intreat you to read it with favour and attention, and to pardon us, wherein we may seem to come short of some words, which we have laboured to interpret. For the same things uttered in Hebrew [ἑβραϊστὶ], and translated into another tongue, have not the same force in them.” (source and translation here.
    — In the Mishna, both the language of the Bible and the spoken language were referred to as לשון הקודש, roughly ‘the holy tongue’.
    — ‘Hebrew’ was used in the Gemara to refer to the old Hebrew script, as distinguished from the later Assyrian script, the ancestor of today’s script.
    — ‘Hebrew’ is used in a few paces to refer to the language in both the Mishna and the Gemara, which suggests to me that the term was generally used, even if the euphemism לשון הקודש was preferred in some contexts.

    To these I add that during the earlier years Haskalah movement of the 19th century and its attendant language revival movement, the language was often called שפת עבר ‘the language of Eber’, an archaism based on עברית. ‘Hebrew’ was used by various Medieval scholars as well.

  7. I should add, Maimonides used the word ‘Hebrew’ (in Judeo-Arabic), in his commentary on the language (Guide to the perplexed 3,8). he argues that the Hebrew language is called ‘the holy language’ because it has no words for the sexual organs, the sexual act, nor for semen, urine or feces, and that the words used for them in the Bible are all euphemisms.

  8. Maimonides lied. The Hebrew word חרא khara shit (not excrement and not feces) appears in II Kings 18:27 (the verse immediately after my cite above) and is repeated in Isaiah 36:12.

    וַיֹּאמֶר אֲלֵיהֶם רַבְשָׁקֵה הַעַל אֲדֹנֶיךָ וְאֵלֶיךָ שְׁלָחַנִי אֲדֹנִי לְדַבֵּר אֶת-הַדְּבָרִים הָאֵלֶּה הֲלֹא עַל-הָאֲנָשִׁים הַיֹּשְׁבִים עַל-הַחֹמָה לֶאֱכֹל אֶת-חריהם וְלִשְׁתּוֹת אֶת-שניהם עִמָּכֶם.

    But Rabshakeh said unto them, Hath my master sent me to thy master, and to thee, to speak these words? hath he not sent me to the men which sit on the wall, that they may eat their own dung, and drink their own piss with you? (KJV)

    When reading these occurrences it is traditional to substitute a euphemism. The word שתן shetten piss/urine is written in both places with the /t/ letter missing.

    Hebrew has only the one word, שתן, to indicate both urine and piss. The word חרא shit has synonyms like צואה tzoah excrement/feces and גללים glalim (animal) droppings in everyday use as well. The baby words — also used by adults — are, respectively, פיפי peepee and קקי kaki, from Yiddish/German.

  9. Paul,

    more on the subject of Sterilizing the Bible.

  10. @bulbul: Thanks. The author quotes the same passage. I’ve bookmarked the site; looks like it carries good stuff.

  11. marie-lucie says

    Thanks for posting that article, bulbul.

  12. Paul,
    I think ‘lied’ is a bit of a strong word. ‘Overlooked’ is probably more apt. The whole essay is, as Maimonides himself says, a digression. He’s just floating an idea. Maimonides was a serious and learned scholar, and has justly been celebrated, but even he could make mistakes.

    There’s more poop elsewhere in II Kings, 6:25 and 10:27.

    These and other errors of Maimonides have already been noted in 1866, in Salomon Munk’s annotated French translation of the Guide to the Perplexed.

  13. Hard to imagine that a man of Maimonides’ intellectual strength ‘overlooked’ these words. In any event, though I can’t read the original Judeo-Arabic,Tel Aviv University had no trouble rendering the passage you cite into Hebrew and even finding a word for ravishing sexual penetration ( שגל shagal) that as it happens appears in Jeremiah 3:2. KVJ translates the term as whoredom.

  14. marie-lucie says

    Do you mean “rape”?

  15. Whoredom renders זנות; been lien with שגל.

  16. Maimonides addresses שגל. He says it derives from שֵׁגַל, a woman kept as a sexual servant (as in Ps. 45:10), and that the word is hence a euphemism. I’m not saying he’s right, just that he does claim that the word is not a counter-example.

    I can’t believe Maimonides was out to deceive anyone by ignoring something written in the Bible, which would have been available and well-studied by his audience. I can easily believe that he overlooked some words, especially since in these days there wre no concordances or dictionaies, and one had to rely on one’s memory. Even today, brilliant linguists frequently make embarrassing mistakes in their languages of expertise. Peer review helps, but was not common in Maimonides’s day.

    Anyway, I mentioned this thing as an early example of the “language has no words/many words for such and such, therefore the language is such and such” genre.

  17. m-l: Do you mean “rape”?

    *I* don’t mean rape, but precisely what Jeremiah or his editors intended is another matter. My sense of the word in that context, and I hasten to note that I’m not a Bible scholar, is that it means something like unbridled lust with both parties consenting — and not necessarily within marriage.

    Towards the end of the verse is a reference to זנות znut, which as MMcM notes, means whoredom. (Hebrew then and Hebrew now have זונה zona for whore; it’s at minimum curious that the word for ‘armed,’ as in ‘we have weapons,’ is מזוין mizuyan. זין zayin in today’s Hebrew is a crude word for the male member, while לזיין li-zayen means ‘to screw.’

    Today’s word for rape is אונס oness (accent on first syllable). Historically the word meant compulsion, and even today is used to describe the Jews of Spain who were forced to convert to Catholicism during the Inquistion: אנוסים anusim.

    The BDB link I provided above for שגל shagal suggests the meaning is “ravish, violate,” and there’s a proviso that the word was considered obscene. The entry immediately after is שגל shegall, whose meaning is given as “(queen-)consort.”

    The Bible *does* steer away from specific descriptions of sexual activity, especially when the activity is condoned. There’s plenty of “he lay down with her” and “he knew her” in those instances.

    Another word, בעל ba’al, has multiple and related meanings. First, it’s the name of a Phoenician/Canaanite deity. An old and still current meaning is ‘owner, master,’ and from ‘master’ we also get ‘master of’ in the sense of ‘skill.’ There’s an infrequently used verb form, לבעול li-v’ol, that means ‘to sexually dominate’. The word is not obscene in itself, but the act it describes is considered so appalling that one does not commonly utter it. I don’t know why it’s not used to mean rape. Yet בעל ba’al also means husband, while the word for wife is ‘merely’ אישה isha woman. The word for female is נקבה nikeva the pierced one . . .

    What was Maimonides getting at? I don’t know. Few are the cultures without difficulty when matters sexual arise (you should forgive the expression). It’s a powerful urge, the source of life (aka dynastic continuity) and the source of much trouble. Perhaps he saw and omitted, perhaps he didn’t see, perhaps he saw and forgot.

  18. Assuming Schniedewind’s dating is correct, it seems like a fascinating observation, and new to me – i.e. that the use of Hebrew as a name for a specific language paralleled sociopolitical changes on the ground. It reminds me of the history of the term “Yiddish,” which is relatively recent as a term for the language used by its speakers (as opposed to “zhargon,” “loshn ashkenaz,” etc.).

  19. Zackary: it seems like a fascinating observation, and new to me – i.e. that the use of Hebrew as a name for a specific language paralleled sociopolitical changes on the ground.

    On the other hand, what is more reasonable to expect than a connection with sociopolitical changes ? Names for languages were not handed down by God to Adam, so far as I know.

    If I understand the quote correctly, Schniedewind is saying that the name “Hebrew” turned up only after “the Judean language ceased to be a living vernacular”, not during its prevalence. Luhmann wrote quite a lot about such historical phenomena, where a name or idea seems to turn up only when the show is over (so to speak). This is a central idea in one of his major works, a multi-volume collection of essays with the title Gesellschaftsstruktur und Semantik.

  20. John Cowan says

    I think that’s intelligible on general grounds. We need names for things to distinguish them from other things: we need names for our own activities when we no longer do them, or when we meet people who do different things. We need the concept of ‘tradition’ only when some of us are no longer traditional.

    For example, among the Romance languages the only places the words Roman or Latin survive for the languages are when they are spoken on non-Romance territory. Latins in Gaul and in Spain needed separate words for their languages to distinguish them, but Romanian and Rumantsch and Ladin and Ladino continued to work fine in their Slavic, Germanic, and Turkish contexts. Similarly, in Navajo diné means either ‘people’ or ‘the Navajo’; the word for ‘lawyer’ means literally ‘people who talk fast and help people’ or else ‘people who talk fast and help the Navajo’, depending on how you see it.

  21. JC: Latins in Gaul and in Spain needed separate words for their languages to distinguish them

    I wonder when that arose? Or more precisely, when did people living in what are today France and Spain — and for that matter in Italy too — decide that the language they spoke was no longer Latin? (I used the verb ‘spoke’ because I assume that most written material in those countries back then was still recognizably Latin — it was the spoken language that had changed, and only later, with more widespread literacy (see: commerce, expanded) did the vernacular come to be written.)

  22. marie-lucie says

    I would think that the names for the peoples came before those of the languages. The name of a language is usually derived from the name of the group that speaks it, itself often derived from the name of their territory. For instance, English started as Englisc, the form of speech used by the Angles, before it became the common language of an enlarged England.

    Even when the vernacular forms of Latin (usually called by a form of romanus or romanice) were still not too far from written Latin, travellers were well aware that the people of Spain, France or Italy did not all speak the same way, even though their respective forms of speech may still have been mutually intelligible. It must have been later that the names of the various peoples were also used for their now separate languages.

    In the case of French, until the Germanic invasions it does not seem that there was a distinctive word for the form of Latin spoken in what was still called by a name continuing Latin Gallia. After the Frankish invaders took over Northern Gaul, learned the local Latin dialect (Rome still being prestigious), passed it on to their descendants but (being now the rulers) imposed on the country their name (hence Francia) and their own way of speaking Latin, eventually causing major differences in pronunciation between the speech of Northern France which evolved into French and that of the areas less deeply influenced politically or linguistically by the former conquerors, by then thoroughly assimilated.

  23. Could shagal have given rise to the British slang word shag, sexual intercourse? It’s used in Canada, too, and a ‘shaggin wagon’ was a van fitted out with a mattress and other comforts.

    And could there be a relationship with the name of Marc Chagall, the painter, who was born Moishe?

  24. John Cowan says

    Shag ‘copulate (with)’ is recorded as early as 1770, so a Hebrew origin is pretty well excluded. In fact, Thomas Jefferson was the author of the first OED quotation: “He had shagged his mother and begotten himself on her body.” (Who, I wonder?) Both the OED and Etymonline agree that it’s < an earlier shag ‘shake’, itself a variant of shake.

    Shake in turn is common Germanic: Etymonline says “No certain cognates outside Germanic, but some suggest a possible connection to Sanskrit khaj ‘to agitate, churn, stir about’, Old Church Slavonic skoku ‘a leap, bound’, Welsh ysgogi ‘move.'”

    As for Chagall, his family name was Segal, supposedly an acronym for segan leviyyah ‘servant to the Levites’. Most acronymic etymologies are bogus, but some Jewish names are an exception.

  25. Thanks, JC, enlightening as usual. But why doesn’t the Wikipedia entry for Marc Chagall refer to Segal, I wonder?

  26. Odd to see my last comment dated in the future. Who says time travel doesn’t exist?

  27. John Cowan says

    LH’s clock is New York time. Chagall’s original name is given in the “Early life and education” section of the WP article.

  28. Who says time travel doesn’t exist?

    Maybe tomorrow it did.

  29. JC: Oops! Caught with my pants down. Didn’t read the entire article.

    PO: Heh! Turning the mind inside out is a healthy exercise.

  30. Semi-serious research for time travel: Top ways to research the internet for time travelers, Earthsky, Jan. 7 2014.

    sorry, I’ve lost my note on how to create a link.

  31. narrowmargin says

    Why is Aramaic referred to as a “Jewish” language? I know that parts of the Bible were written in it but wasn’t it simply the lingua franca in the Levant 2,000 years ago, unattached to any religious or ethnic group? What am I missing here?

  32. John Cowan says

    It was, like Akkadian before it and Arabic after it. But also like those, it became the native language of a number of different populations, including the Judaeans. Consequently, not only parts of the Bible but the bulk of the Talmud are written in it, and so it is widely studied by Orthodox Jews.

    Small groups of Jews and Christians, as well as Mandeans, continue to use it as a first language, often in conjunction with Modern Hebrew or Arabic. The varieties are often no longer mutually intelligible, and Ethnologue identifies 14 separate languages still in use. Only two of them are considered “Vigorous”, defined by Ethnologue as “used for face-to-face communication by all generations and the situation is sustainable”, but without literary form or standardization.

  33. narrowmargin says


  34. The First Nation Cree of North America speak a Biblical Hebrew. Its is called NayheYHWHwin. It is the only Language with the Holy Name of God within it. The Ojibway known as Anishnabi also speak this biblical Hebrew.
    Wgy isn’t there any etymological studies for the First Nations of Canada? We the Cree, Anishnabi speak the Ancient languages of the east. We also have Arabic and Ugarit as well including other from the middle east too numerous to mention. We the First Nations can fill the gap between expectations and results. Btw most if not all First nations have our languages still intact. The residential schools failed in trying to eradicate our languages.

  35. Michael Vinegrad says

    @Paul Ogden, @bulbul: the above link for Sterilizing the Bible is broken. The article’s URL is https://marginalia.lareviewofbooks.org/sterilizing-the-bible/.

  36. Examples of foul language in Tanakh overlooked by Maimonides, which were mentioned upthread (2 King 6:25, 10:27,18:27, Isaiah 36:12) are all examples of ketiv/qere that is they are read aloud differently from the standard text (and it is written into the text itself). That makes it more unlikely that the big M simply forgot about them and somewhat likely that he considered them euphemized away. Because I haven’t read Rambam I cannot say whether it is a plausible explanation.

  37. λάλησον δὴ πρὸς τοὺς παῖδάς σου Συριστί

    Adding παῖς to the collection of words that can mean both “child” and “servant”. And it goes back farther than the LXX; e.g., Liddell-Scott-Jones cites a scene in Aeschylus where Orestes calls for a servant with “παῖ παῖ”, rendered in one English translation by “Boy! Boy!”

  38. David Eddyshaw says

    Persicos odi, puer, apparatus …

    Incidentally, the story I imbibed in my youth that puer was the second element of the slave name “Marcipor” (etc), and that this meant “Marcus’ boy/slave” seems to have been too simple, or rather only half right:


  39. I picked up the book that inspired by the thread, and I’m sure I’ll learn from it, but this has me worried that I may need to be more skeptical than usual.:

    >The writing system was invented by the Sumerians and borrowed for the Akkadian language, even though Sumerian is an Indo-European language and Akkadian is Semitic.

  40. The discussion of Sumerian-IE connections in WP is not too detailed: “Sumerian was at one time widely held to be an Indo-European language, but that view later came to be almost universally rejected.[16]
    Among its proposed linguistic affiliates are:

    Then they list everything without mentioning IE.

  41. Anyway, if the hypothesis was once popular*, maybe it is one of those “facts” that once were fashionable and keep migrating from textbook to textbook after they ceased to be so.

    *WP’s reference for this is Dewart, Leslie (1989). Evolution and Consciousness: The Role of Speech in the Origin and Development of Human Nature. p. 260. that has a footnote:

    2. At one time it was widely held that Sumerian was among the IE languages; cf. C. Autran, Sumérien et Indo-Européen: l’aspect morphologique de la question (Paris, 1925). Currently the almost unanimous opinion of linguists holds that this is most unlikely, and that Sumerian should remain unclassified as to family until further evidence should appear. I shall assume here that this view is correct; I know of no theoretical reason to the contrary, and in any event the truth of the matter has no bearing on present purposes. However, the possibility of common ancestry, going back to the pre-thematic stage of human evolution, cannot be discarded simply because Sumerian is not an IE language.

  42. What do you think Leslie Dewart (he/him) meant by “the pre-thematic stage of human evolution”?

    Sure there was a time in the 20th century when Sumerian was believed to be IE. But this is a prof of Semitic languages writing in 2013, who didn’t even make room for the possibility that it might not be. And edited by a university press, who apparently didn’t notice.

  43. David Eddyshaw says


    This appears to be a Heideggerian technical term, though I hesitate to accuse a respectable scholar of Heideggerian tendencies. As far as I can make out, it means something along the lines of “intuitive; before rational thought comes along with its pesky urge to categorise and spoils everything.”

    Presumably Dewart was thinking of an age before humanity became rational (at all.) I would agree that in that case, minor details like the actual affiliation of Sumerian would become quite irrelevant, as indeed would a great deal else of what we actually know about language …

  44. The book is here: https://archive.org/details/evolutionconscio0000dewa/page/260/

    It is borrowable, and also I can see this page without borrowing it… Oh! I can even see the next page by changing manually 260 to 262 in the URL.

  45. David Eddyshaw says

    It does not inspire confidence that the very first sentence on that page is

    Indeed, ‘to be’ and like copulative verbs are exceedingly rare; only the languages of the Indo-European family, and outside these only Sumerian – the language of the first human civilization, which emerged in Mesopotamia about 5,500 years ago – can be confidently attributed this feature.

    As the Kusaasi say:

    Dau kaŋa nwa anɛ zɔlʋg.
    Man demonstrative.SG this COPULA.FOCUS fool.SG.

    I presume that the Kusaal copula is of Sumerian origin.
    This is not the writing of one who believes in checking his facts.

    Sumer, of course, was not the first human civilisation (though it was probably the first literate civilisation.)

  46. Stu Clayton says

    As far as I can make out, [“pre-thematic”] means something along the lines of “intuitive; before rational thought comes along with its pesky urge to categorise and spoils everything.”

    “Intuitive” yes, but a little more sharply: vorthematisch suggests “before X was/could-have-been a notion to waffle about, even intuitively”. At this stage, questions about how to categorize it are even more inconceivable.

    vorthematisch in this sense is a word I would expect to find in Husserl, and even earlier.

    It’s all part of a strategy to talk about stuff more or less, but without hypostasizing it (in contrast to “Look, a squirrel!”). Kant today would have to startle the world with das Vorthematische-an-sich. You can’t ding anyone with a Ding anymore.

  47. It does not inspire confidence that the very first sentence on that page is
    See the note three, though. He distinguishes copulas that “reflect objective reality” from copulas that do not. Apparently, a word that combines copulative and existential meanings, indicates for him this objective interpretation.

  48. David Eddyshaw says

    Ta, Stu; I thought you would be likely to know, if anyone.

    Dewart does indeed seem to have been well into the phenomenology.

  49. David Marjanović says

    …OK, I admit I haven’t even peeked into the grammar of Kusaal (though I think I’ve downloaded it…?). Is the copula a verb?

    Modern Mandarin has a copula, but the very un-IE ways in which it is used or rather not used betray its origin from a demonstrative rather than a verb.

  50. David Eddyshaw says


    He seems to groping (with much obfuscation) towards the familiar-enough assertion that the conflation of the existence-verb and the copula in Greek and Latin led to philosophical confusion. Presumably speakers of (Indo-European) Spanish and Irish are free of such confusions. (And Welsh speakers, but only so long as they abstain from positive assertions.)

    I note that he uses “objective reality” both of the reality apparently signified by an existence verb (as in “unicorns exist”, “an ongoing process of obfuscation exists here”, “there is no help for it”) and the reality of a “relationship” expressed by the copula in an indo-European language, as in “these roses are red”, where there is presumably some sort of objectively “real” connection, a bit like a piece of string, perhaps, between “these roses” and “red.” This is a pitiful confusion of form and function, but as the thesis of the book is apparently that consciousness is (somehow) generated by language, I suppose he was committed to this confusion ex hypothesi.

    My own feeling is that Dau kaŋa nwa anɛ zɔlʋg is a perfectly objective Kusaal statement, even though Kusaal has a different verb for “exist” from “be something.”

  51. Stu Clayton says

    @DE: For relaxation, I occasionally read a few pages of one of these modern guys with a “phenomenological” style, like Hartmut Rosa. I drift along with the wift, gathering woolly impressions of epistemological thisses and thats. Here are some of his book titles:

    Alienation and Acceleration. Towards a Critical Theory of Late-Modern Temporality
    Unverfügbarkeit (Unruhe bewahren)

    But excelsior ! This year he and Andreas Reckwitz published Spätmoderne in der Krise: Was leistet die Gesellschaftstheorie ? In it they ignore Luhmann completely, which is just unprofessional in a book on Gesellschaftstheorie – even when the book is intended for the bushy-tailed cognitive bourgeoisie, as this one is.

    Luhmann is indeed not easy reading – but Wenn man das Pfeifen und das Trommeln nicht verträgt, soll man nicht in den Krieg ziehen”.

    40 years ago Luhmann gave acidulous descriptions of naive, earnest poppets who fawn on the “fathers” such as Weber and Durkheim. They’re still with us (the poppets).

  52. David Eddyshaw says

    Is the [Kusaal] copula a verb?

    Yes, undoubtedly. It’s a well-behaved member of a whole minor conjugation of verbs found only in the imperfective aspect (Kusaal has about sixty of them altogether.)

    You can use it in the past:

    Dau kaŋa nwa daa anɛ zɔlʋg “This man was a fool.”

    You can use it in direct commands:

    An baanlim! “Be quiet!”

    It has a gerund anlim “being something”, and my chief informant even happily produced an agent noun for it without hesitation: aand “someone who characteristically is something.”

    The form anɛ is due to a confusion in the traditional orthography: the is a focus particle, which turns up in many other contexts as well; it’s usually written solid with the preceding verb when it’s been recognised as the focus particle (there are two other quite distinct particles to confuse matters.) The verb itself is really aen [ãj]; it loses the [j] when not phrase-final, by a general sandhi rule applying to all verbs, and loses the nasalisation before by another sandhi rule which also applies generally, not just to this verb.

    It has more in the way of morphology than is immediately apparent, concealed by the pervasive Kusaal deletion of final short vowels in citation forms (and indeed in most other contexts too.)
    In e.g.

    Amaa fʋn pʋ aenya “But you’re not.”

    it reclaims its full untruncated form [ãja]: the -ya flexion revealed here is characteristic of the whole minor conjugation.

  53. Modern Mandarin has a copula, but the very un-IE ways in which it is used or rather not used betray its origin from a demonstrative rather than a verb
    There have been speculations that PIE *h1es- was originally a demonstrative as well. IIRC I first found that idea in Hirt’s IE grammar, but I think I’ve seen that idea expressed elsewhere, too.
    Presumably speakers of (Indo-European) Spanish and Irish are free of such confusions. (And Welsh speakers, but only so long as they abstain from positive assertions.)
    Positive assertions only bring trouble.

  54. Positive assertions only bring trouble.

    I just read this piquant passage in James Billington’s magisterial The Icon and the Axe (p. 160):

    The final years of the seventeenth century were dominated by more negative protests against the new order, reaching a climax in the movement to abjure all worldly speech save repetition of the word “no”—the famous netovshchina of a peasant from Yaroslavl named Kozma Andreev.

    Alas, the Russian Wikipedia article on the movement does not support that suggestion of extreme laconism. It’s like finding out the boring truth about the Circumcellions.

  55. Stu Clayton says

    Ich bin der Geist, der stets verneint!
    Und das mit Recht; denn alles, was entsteht,
    Ist wert, daß es zugrunde geht;
    Drum besser wär’s, daß nichts entstünde.

  56. David Eddyshaw says

    To be scrupulously fair (when am I not?) there is something fishy about the copula in Western Oti-Volta, not in Kusaal itself, where it is a perfectly respectable one-aspect verb among many others, but from a comparative standpoint. It doesn’t seem possible to reconstruct a copula even for Proto-Western Oti-Volta, let alone any farther back; unlike the “exist” verb (Kusaal ) which goes way back, certainly beyond Proto-Oti-Volta and probably even to Proto-Volta-Congo.

    Moreover, the copulas of most other Western Oti-Volta languages are often much less like proper verbs. In Mooré, for example, the copula yaa need not have a subject expressed, even though Mooré (like Kusaal) is not “pro-drop” (ugh):

    Mooré: Yaa sõama. “It’s good.”
    Kusaal: Li an sʋm. (where li “it” cannot be omitted)

    The various copulas across WOV do look rather like different 3rd person pronouns, but that probably doesn’t mean much with such short morphemes, especially as the legacy of noun-class-based grammatical gender is that there are a lot of different potential 3rd person pronoun forms to play pick and mix with.

  57. David Marjanović says

    There have been speculations that PIE *h1es- was originally a demonstrative as well.

    Well, *h₁e- is a known demonstrative (Latin is, German er…), so if the “root extension *-s- which is found in a lot of verbs was somehow added to turn it into a verb, then maybe…

    But with such short forms that strikes me as rather untestable. Indeed I’m not aware of evidence for *h₁ in the demonstrative other than the general hypothesis that PIE didn’t allow vowel-initial words.

    (That’s not the case for the verb, which preserves evidence for *h₁ in Hittite, Greek and IIRC Sanskrit.)

  58. David Eddyshaw says

    I was wondering if at least some of the copulas in Western Oti-Volta might be derived from pronouns coopted into the verbal system.

    Most are non-starters: the Mampruli copula is nye, which, if it’s related to anything at all in Kusaal, would presumably be nyɛ “see, find”; Mooré yaa doesn’t really look like any pronoun at all; Dagara has ʔi, which would do for a plural pronoun but would belong to the class with the least members of all ….

    Boulba has , which is negated like a verb, with the same preverbal negative particle, but looks very like a repurposing of the word which turns up in Kusaal as nwa “this” (which may itself have originated as a by-form of the 3rd sg human-class pronoun.) So that’s a candidate, though I can’t find evidence in the very scanty materials available that it has any distinctively verbal morphology.

    Kusaal and Farefare both have copulas that are unequivocally minor-conjugation verbs, but they are, unhelpfully, clearly not cognates: Farefare has de, which resembles the singular pronoun of the single largest noun class: di (= Kusaal li.) The Farefare copula does have distinctively verbal flexion, using dɛna for imperative and consecutive forms, in keeping with the minor-conjugation pattern in Farefare.

    The Kusaal copula aen goes back to *ŋaya, where the -ya element is the minor-conjugation imperfective flexion; the root *ŋa looks like *ŋa, the 3rd plural pronoun of the single largest noun class.

    If the roots of the Farefare and Kusaal copulas do go back to pronouns, in both cases they have been coopted into small “irregular” verb conjugations. This is not quite so implausible as it may seem: Kusaal, for example, has a verb labi “lurk behind something in hiding to eavesdrop” in the minor conjugation, which is surely connected with the synonymous Hausa verb laɓe, in view of the coincidence of both form and highly specific meaning. There is no question of Hausa having borrowed the word from Kusaal, and the Kusaal word seems to have no cognates elsewhere in WOV, so this looks like a definite loanword. The minor conjugation has a fairly clearcut set of meanings associated with it (qualities, relationships, bodily stances and positions), so even though it has few members, it’s perhaps not such a stretch that it might recruit the odd loanword (and in this particular case, the analogy of vabi “lie prone” might well have helped.)

    I think, putting all this together, the most plausible scenario is that Proto-WOV didn’t have a copula at all, and that the various WOV branches have independently innovated one, in some cases (perhaps) from 3rd person pronouns, but certainly not always (if at all.) I’m not sure quite what would have driven this independent innovation of copulas, though. (Welsh substratum …?)

  59. David Eddyshaw says

    Pursuing my line of thought (you were warned) …

    Kusaal minor-conjugation verbs are often derived from adjectives by adding *-ya to the adjective stem (the communis opinio seems to be that adjectival verbs in Oti-Volta are primary, but the tonal evidence in Kusaal refutes this, at least for Kusaal, and presumably the rest of WOV.)
    Thus e.g. tɔɔg (*tɔɔgu) “bitter”, tɔe (*tɔɔya) “be bitter”; giŋ “short” (*gimga), gim (*gimya) “be short” etc.

    One or two derive from human-reference nouns e.g. kpɛɛnm (*kpɛ̃ɛ̃mya) “be older than” from kpɛɛnm (*kpɛ̃ɛ̃mu) “elder (person.)”

    Conceivably, the same *-ya suffix could once have been added to pronouns, creating verbs meaning “be a thing/person of the relevant noun class.” If such verbs existed, it would be natural enough for them to develop into copulas; they would then (effectively) agree in class and number with their subjects.

    This sort of agreement, however, would be extremely exceptional in these languages. It would cry out to be reduced and simplified, not least because most of the languages have lost grammatical gender altogether. You might, as an intermediate stage, end up with a sg/pl pair of copula verbs, say *liya (or *diya) and *ŋaya. Even this would be a great anomaly in languages where verbs show no agreement of any kind; on first principles, you might expect the singular form to prevail at the expense of the plural (et voilà! Farefare!); however, in Kusaal, with its loss of final vowels, this singular form would end up absolutely homophonous with either li “fall” or di “eat”, so there might be pressure to prefer the originally-plural copula instead.

    Kusaal, Farefare and Boulba are the only languages which have preserved the minor conjugation as a recognisable entity; other Western Oti-Volta languages have (it seems) nothing more than stray relics or nothing at all. Any of these hypothetical copula verbs would thus get stranded, with no morphological group to belong to, and end up as invariant particles resembling old pronoun forms – or get replaced altogether by something more transparent.

  60. Sure there was a time in the 20th century when Sumerian was believed to be IE.

    I did not know:( I was trying to find something about it.

    But this is a prof of Semitic languages writing in 2013, who didn’t even make room for the possibility that it might not be. And edited by a university press, who apparently didn’t notice.

    Yes, he was sloppy. Perhaps I should not expect this* from a Biblical scholar, but he did not check.

    I just was thinking why he was so confident in it. People are often inaccurate with neighbouring fields and there are “facts” reproduced in textbooks for centuries, because the writer read it in a textbook herself and is not really interested in the matter.

    *”this” – knowing the current ideas about Sumerian.

  61. Koenraad Elst, Some unlikely tentacles of early Indo-European:

    …From the Indian viewpoint, these scenarios all imply that the Indo-Aryan branch of the IE family was brought in from outside by “Aryan invaders”. Therefore, among Indians and India-watchers, the different versions of European or West-Asian Homeland hypotheses are jointly called the Aryan Invasion Theory (AIT).

    The Indian answer to the AIT has been called the Out-of-India Theory, a term which to our knowledge was coined by US scholar Edwin Bryant ca.1996, when he was writing his Ph.D. dissertation on this dispute. He assumed the kinship and (at least partial) common origin of the Indo-European languages as a given, with the implication that either its non-Indian branches originated in (and expanded or emigrated from) a Homeland in India or the Indian branch(es) moved into India from a non-Indian Homeland; so denying the AIT automatically implied positing the OIT. But this describes the position of only a section, or rather two sections, of Indian anti-AIT opinion. On the one hand, a non-academic group of Hindu nationalists, starring the late P.N. Oak, assert that more or less everything of value in the world was brought or created by ancient emigrants from India. Thus, England really is the Hindu colony Angulisthân, the Vatican was founded as Veda-Vâtikâ,and Dutch is the language of the Daitya-s. We need not discuss that school of thought in any more detail here. On the other hand, a very small handful of scholars, esp. Misra (2005) and Talageri (2000, 2009), have tried in earnest to build a case for a scenario in which emigration from an Indian Homeland explains the presence of IE languages outside India.

    However, most Indian scholars who oppose the AIT, both professional and amateur, including hundreds of archaeologists, confine their effort to refuting the Aryan invasion of India. Their horizon stops at the Khyber Pass, questions about the IE languages outside the subcontinent simply aren’t an issue for them. This Neither-Immigration-nor-Emigration position is definitely the most popular answer to the AIT among Indians (e.g. Lal 2008, Kalyanaraman 2010). When pressed for an explanation for the commonalities between Indo-Aryan and the European languages, they tend to insist that mutual borrowing between otherwise unrelated languages can take care of the unmistakable similarities (e.g. Waradpande 1989). Whoever has studied the linguistic data and method more closely than they have, cannot follow them down that road.

    I am familiar with this attitude.

  62. David Marjanović says

    this singular form would end up absolutely homophonous with either li “fall” or di “eat”

    Man ist, was man isst.

    Admittedly, ist has lost its /t/ in practically every dia- and mesolect, while isst never seems to do that.

  63. works, more or less, in Russian as well.
    человек есть то, что он ест

  64. David Eddyshaw says

    Yes, I’m not sold on this theory myself, by any means. The weakest link in it is surely explaining why an originally plural copula would have got generalised at the expense of the singular, which seems pretty implausible on the face of it.

    It is very odd that such closely related languages should have copulas that can’t be directly related to each other, though.

    It occurs to me that these hypothetical multiple copulas would in fact have “agreed”, not with the subject, but with the complement, and that this would not always be a noun in any case. Kusaal actually rarely uses adjectives as complements of the copula (though this is grammatically possible): the usual way to say “it’s good” in Kusaal, for example, is either

    Li an sʋŋa. “It is well.” (adverb derived from the adjective with -ga.)
    Li an sʋm. “It is goodness.” (abstract noun created by inflecting the adjective in the mass/abstract class -m.)

    [You could also use the corresponding verb and just say Li sʋm. “It is-good.”]

    Manner-adverbs are a kind of abstract noun in Kusaal syntactically, so these cases are not as different as they might appear.

    So: the copula that got generalised in Kusaal (in this scenario) could actually be the one that “agreed” with adverbs (“be somehow”), whereas Farefare would have generalised the one for nouns (“be something.”)

  65. If “it” is abstract, we use adverbs and (occasionaly) abstract nouns with it in Russian as well…

  66. Stu Clayton says

    Yes, he was sloppy. Perhaps I should not expect this* from a Biblical scholar, but he did not check.
    I just was thinking why he was so confident in it. People are often inaccurate with neighbouring fields and there are “facts” reproduced in textbooks for centuries, because the writer read it in a textbook herself and is not really interested in the matter.

    To call “inaccurate” what for centuries were taken as “facts” is anachronistic. Only now are you in a position to regard them as “not facts”. When people take A to be a fact, they do so in all confidence until its (unforeseeable) fact-until date T expires (when A is replaced by B).

    “Fact” is an energy-saving cognitive honorific. It is used to encourage others to accept what it refers to – not arbitrarily, of course, but for reasons that appear to be good. Accepting at some point that A is a “fact” means one can in good conscience cease thinking and quarreling about it, and move on to think and quarrel about other topics. “Not really interested” in A then is the right way to put it: why be interested, when one has license to be disinterested ?

    Of course sometimes people lose their license and, in order to scrape by, fall to writing exciting, fascinating books on the facts about UFOs and aliens.

  67. To call “inaccurate” what for centuries were taken as “facts” is anachronistic.

    There may be something in that thought, but it’s irrelevant to the discussion. Sumerian was not believed to be IE “for centuries,” it was thought by some to be so for a bit in the early 20th century. And Schniedewind was not writing back then, his book came out in 2013, by which time it took a striking degree of ignorance and indifference to perpetrate the musty error.

  68. Stu Clayton says

    I merely quoted drasvi, piling words on words. Those are the facts here. You can take up with them the question of how could they say such a thing.

    By the way, there’s a fancy German expression describing the human condition of “the backward and the forward living side by side”: die Gleichzeitigkeit des Ungleichzeitigen.

  69. There are no facts involved, just verbiage. Of course, I realize that your point of view does not allow for “facts” in the first place, but still. Drasvi is making a point about scholarship, you are making a point about… something, but it doesn’t have anything to do with what you quoted as far as I can see. Which is fine, it’s your thing, I just wanted to set the record straight.

  70. J.W. Brewer says

    The Schniedewind blunder (and/or the Yale University Press blunder during the editing process …) is indeed quite peculiar. Schniedewind has co-written “A Primer on Ugaritic,” so he knows at least one ancient tongue beyond Hebrew but I wonder if he just conceptually divides all ancient Near Eastern languages into “Semitic” and “not-Semitic,” has no particular need to keep track of variations and affiliations among the non-Semitic ones since he is himself a Semiticist, and, I don’t know, got Sumerian muddled up in his head with Hittite or something. In terms of the offending sentence, the important part from his POV may well have more likely been “Semitic language borrowed writing system from non-Semitic language” than “Indo-European language’s writing system was borrowed by a non-IE language.”

    I also learn from the internet, FWIW, that “Schniedewind was the director of the Qumran Visualization Project (QVP), which created a virtual reality model of ancient Qumran under the auspices of UCLA’s Experiential Technologies Center.”

  71. Stu Clayton says

    “I have nothing against facts, but I wouldn’t want my daughter to marry one.”

  72. Apropos of nothing, here’s a striking example of academic nerdview (from Timothy K. Blauvelt, “Clientelism and Nationality in an Early Soviet Fiefdom: The Trials of Nestor Lakoba”):

    The relationship between the central Soviet leadership and the local national elites often resembled that of a grantor with a grantee: before the finalist selection has been made, the grantor has all the power and can make the applicants jump through hoops; once the choice has been made, however, and the grant awarded to one of the applicants, now the success of the grantor depends on the success of the grantee. This alters the power relationship, allowing the grantee to make demands on the grantor: a kind of “capture” results. In a similar way, the central leaders need to make a choice of on whom among the local titular elite to place their bet, and once they have chosen, the patrons to an extent become dependent on the clients.

    A grantor and a grantee! Who outside the inbred world of academia would be likely to understand that analogy?

  73. J.W. Brewer says

    The pages around that blunder about Sumerian, fwiw, do contain shout-outs to multiple famous linguists, like de Sassure, Bloomfield, and Labov, and perhaps Schneidewind ought to be praised for the fact that (assuming search results are complete …) the only Chomsky referenced in the book is William — Noam’s father, but more to the point the author of a 1951 article which sounds from its title perfectly relevant to the topic of Schneidewind’s book.

  74. Yes, I’m prepared to forgive him much for that.

  75. Stu Clayton says

    I understood the paragraph without knowing what “grantor” and “grantee” mean in whatever context they’re taken from. I just worked with the pattern V+”or/ee” for any verb V. Is “grantor/grantee” really a pre-thematic thing ? [I’m trying to wear out the word “pre-thematic”]

  76. But it’s not a useful analogy if the reader has no prior understanding of the grantor/grantee relationship. You can understand the relationship between gostak and doshes from the phrase “the gostak distims the doshes,” but it doesn’t do you much good.

  77. Stu Clayton says

    Is there a particular reason why you resist taking it as an instance of applying the pattern V+”or/ee” for any verb V ? People make up word pairs in that way all the time. If you know what “grant” means, you know what “grantor” and “grantee” mean.

    Another example: “allegator”/”allegatee”. The allegator makes an allegation against the allegatee. Usually on the banks of the Nile, but this is not required.

  78. Is there a particular reason why you resist taking it as an instance of applying the pattern V+”or/ee” for any verb V ?

    Yes, because that tells you nothing about the particulars of the relationship and the people to whom it applies. Words have a purpose beyond supplying formal patterns, you know.

  79. Stu Clayton says

    The relationship is of one who grants, to one who receives the grant. What could be simpler ?

    As to words having a purpose beyond supplying formal patterns, I find that suggestion pretty extravagant. For instance, the exchange of formal patterns we are now engaged in seems to have no purpose. Surely we can agree on that at least ?

  80. The relationship is of one who grants, to one who receives the grant. What could be simpler ?

    Nothing could be simpler, or more meaningless. Are you just screwing around, or do you not understand why that’s useless for understanding the analogy?

  81. David Marjanović says

    Who outside the inbred world of academia would be likely to understand that analogy?

    That’s why he explains it.

    Making up grantor & grantee, which I’ve never actually encountered before, even makes it more transparent than funder and PI (Principal Investigator, the person who actually submits the grant proposal on behalf of however many people).

  82. Stu Clayton says

    The paragraph you quoted describes a standard vanilla Herr/Knecht relationship. You find it everywhere: person A promotes person B to serve the interests of A, after vetting person B. A grants B a special position. Then A is obliged to support the actions of B, because B was A’s choice. A becomes dependent on B.

    This is why the GOP finds it so hard to deal with Trump.

    Example from my own experience: at the McDonald’s where I worked, the franchiser appoints a guy to be his supervisor in the store. This guy turns out to be a serious asshole on a daily basis, but the franchiser will not acknowledge any wrongdoing. Much later, after I had quit, the franchiser explained to me that when you appoint someone to represent you, you have to support him 100% or fire him. Otherwise you undermine your own position.

    ETA: I gather from David’s comment that you both may be thinking of something like research grants. But I wasn’t. As I wrote: “I understood the paragraph without knowing what ‘grantor’ and ‘grantee’ mean in whatever context they’re taken from.” I even asked, in a jocular way: “Is ‘grantor/grantee’ really a pre-thematic thing ?”, but got no answer.

    I simply took the everyday word “grant” to mean what it does.

  83. I’m with Stu here – I didn’t even think of research grants when I read the sentence in your post. One can grant many things, among them favours and titles. The latter was what came to my mind immediately, especially in the context described.

  84. Well, *h₁e- is a known demonstrative (Latin is, German er…), so if the “root extension *-s- which is found in a lot of verbs was somehow added to turn it into a verb, then maybe…

    Or you could combine it with the other demonstrative *s-. Which would be kind of neat, but yeah, pretty much untestable.

    Incidentally, after twenty-odd years of studying Greek I recently discovered a personal pronoun heretofore unknown to me: ἵ, apparently only preserved by a couple of grammarians and meaning either “he, she” or just “she”. (Greek will do that). It seems to be the nominative of the third-person pronoun οὗ (gen.) οἷ (dat.) ἕ (acc.), which grammars usually list as lacking a nominative, and may be cognate with Gothic si and Irish , both “she”. If so, and if that somehow comes from *(h₁)si (based on the masc. nom. sg. *(h₁)is), that would argue against a laryngeal in that demonstrative, but the second if seems particularly unlikely.

  85. @David Marjanović: Some of the “grant” terminology is specific to funding in the arts and humanities. The money itself is a “grant,” but I wouldn’t normally describe myself (or my institution) as “grantee” for the funded projects on which I have been principal investigator or co-PI. And I would never refer to the National Science Foundation or the National Institutes of Health as the “grantor.”

    To the extent that the relationship being described is not totally generic,* the nature of the ​relationship between funding agency and funded individual is probably more typical of the situation with grants from the National Endowment for the Humanities or, especially, the National Endowment for the Arts. If an artist gets support to produce a work, there’s not much that can be done to set intermediate benchmarks or to evaluate whether the money is being spent on a productive development process. Moreover, to some extent, the point of funding creative endeavors is just to give a creator the freedom to create, without immediate concern about economic success. Norton Juster created his magnum opus while he was supposed to be working on a federally funded book about the nature of cities. He recognized that wasn’t making much progress, so to keep busy, he started collaborating with his neighbor, the cartoonist Jules Pfeiffer, on an illustrated children’s book, which became The Phantom Tollbooth.

    * Whenever one party invests money in another’s project, the first party acquires a measure of dependency on the second. How much dependency is determined by the sums involved. There are adages about how, if you owe the bank a hundred thousand dollars and can’t pay, you have a serious problem; but if you owe the bank a hundred million dollars and can’t pay, the bank had a serious problem

  86. David Marjanović says


    Pages 22–23 of this paper set up a PIE system of, oddly, stressed pronoun stems in *-í- and adverbs in *-e:

    *kʲí- “this here” – *kʲe “here (to)”
    *dí- “that over there” – *de “there (to)”
    *í- “that mentioned before” – *e “earlier”
    *(h₁)rí- (hypothetical) – *(h₁)re “back” (only Italic and Gothic)

    No Greek reflex is listed for *í-. But if that was *h₁í-, it should have become hi- in Greek by Bozzone’s law. The h in the other cases could even be from *h₁j- by the same law.

    It’s just odd that the endingless form is supposed to be the nom. sg..

    I can in principle imagine that *h₁es- could then be from *h₁e “earlier” through a meaning “has always been”, “is statically”. But of course that’s a slippery slope to *sal, *ber, *jon, *roš, *kat.

  87. It’s just odd that the endingless form is supposed to be the nom. sg..

    That’s why the Greek form has been compared with the Gothic and Irish si “she” forms I quoted, though the origin of those seems to be unclear. IIRC the grammarians who mention ἵ don’t specify its gender and list it with οὗ οἷ ἕ which are both masc. and fem., but in the one fragment of Sophocles where it appears the referent seems to be feminine. Whether and how it’s related to other weird pronominal forms with -i- like μιν / νιν (both acc.) or dialectal ἵν / ϝιν (apparently a dative!) is a mystery to me.

  88. David Eddyshaw says

    Gothic and Irish si “she”

    (and Welsh hi …)

  89. Nah, that’s clearly a relic of the Hebrew substrate.

  90. David Eddyshaw says

    Of course. Silly of me.

    Of course, the Greek change of *s -> h is, in turn, due to the Welsh substrate.

  91. David Marjanović says

    The si-type ones could be from the demonstrative stem *so- with the feminine ending *-ih₂ (both in zero-grade, which, I guess, should be possible if the whole word was unstressed enough), followed by vowel shortening in sufficiently unstressed words as needed – as long as we can keep Greek out. The Greek outcome of *ih₂ is famously not *ī, but *ja.

  92. David Eddyshaw says

    Welsh hi, at any rate, must go back to a form with a long vowel; and the Old Irish is in fact (scrolling back, I see that TR actually said so.)

    The Gothic presumably has a short vowel, but then, Germanic is weird. I blame the Semitic substratum …

    (Incidentally, I’ve just seen that Joseph Wright actually cites that very Greek form when discussing the etymology of the Gothic personal pronouns. There ain’t half been some clever bastards, as Ian Dury rightly reminds us.)

  93. Weiss in this article (where I first learned of ἵ) thinks the vowel was originally short, fn. 21.

  94. David Eddyshaw says

    Latin too!
    Mind you, it looks like people have read almost anything into that text …


    I don’t know if the long vowel underlying Welsh hi can be explained as “a short vowel lengthened in absolute final position”, as Weiss suggests for the Irish; but there’s mi “i”, I suppose. So yes, could be. Though now we’re talking

    Originally long (see DM above)
    Shortened due to lack of stress (including in Greek, which doesn’t seem to do that sort of thing much, unlike Germanic; nor does Celtic, AFAIK)
    Lengthened again

  95. David Eddyshaw says

    Lengthening of short vowels in absolute final position is certainly a thing in Old Irish, whatever else.

    The infixed (i.e sandwiched between preverbal particle and verb) 3rd person pronouns in Welsh (Middle or Extremely Literary) are -i sg -u pl (which are really just variants of the ordinary proclitic pronouns) or -s for both genders and both numbers; the choice is determined only by which preverbal particle is used. No idea how that came about. The -s- would have to go back to -ss-, or some other such cluster, too. Old Irish infixed pronouns look marginally more comprehensible … though, this being Old Irish, also much more complicated.

  96. Sumerian was not believed to be IE “for centuries,” it was thought by some to be so for a bit in the early 20th century.

    Possibly, Sumerian has “been believed to be an IE lanaguage” for some 100 years within a community of Biblical scholars. I am just aware of the mechanism: a professor X shares a fact in an introductory course. His student professor Y shares the same fact in her introductory course (because it was in the introductory course, it must be somethign basic). This chain of transmission can continue for a long time. I think it is not difficult to find examples that actually travel from book to book for centuries. What I was thinking about is 19th century stuff. Say, that story about people who can’t count beyond “one, two, many”.

    And Schniedewind was not writing back then, his book came out in 2013, by which time it took a striking degree of ignorance and indifference to perpetrate the musty error.

    I disagree about “ignorance”. Ladies blush when you misclassify Russian in a polite society, but not when you do it to Lingala.

    Not many people can actually use the status of Sumerian in analysis of Rabbinic literature. “We do not know what Sumerian is”. Use it:/ Even though Sumerian philology as practiced by Mesopotamian Semites has to do with the Hebrew world, how much does a specialist in Greek literature know about Egyptian?

    But I agree with Ryan.

  97. PlasticPaddy says

    Monosyllables ending in a vowel in Irish are a little strange. For example, te “warm/hot” has two variants in Middle Irish, one with short e like the modern word and one with long e. Matasovic has Proto-Celtic *tefent for this but also *tefnet for the Modern Irish tine “fire”. The latter has preserved the second syllable and the n. I am sute there is an explanation involving replacement of n with nasalisation, followed later by restorative replacement of nasal V with Vn, but I still find it strange.

  98. @Stu, the theory of “disinterested” needs elaboration. Why do disinterested people want to share a fact?

    Are not people who share facts about aliens exactly disinterested? Interested people are supposed to 1. admit that they do not know a shit about aliens 2. speculate 3. quarrel (I am not sure about 3. but you mentioned it).

  99. Trond Engen says

    Heh. With a morpheme (or a conjugation) that derives verbs meaning “be [something]” from adjectives and suchlike, the need for a separate copula is greatly reduced. The bleaching of a derived verb to a new copula might have gone hand in hand with the verbal derivation becoming obsolete or restricted.

    The derivational morpheme could itself be a cliticized copula. Jespersen cycles all the way down.

  100. David Eddyshaw says

    With a morpheme (or a conjugation) that derives verbs meaning “be [something]” from adjectives and suchlike, the need for a separate copula is greatly reduced

    True; and the use of adjectives as complements of the copula even now in Kusaal is quite limited; outside Western Oti-Volta, this seems to be even more the case, with quality verbs pretty much always used instead of predicative adjectives, although the verbs are actually usually held to be primary, and the “adjectives” to be really cases of infixation of verbal stems between noun stems and their noun class suffixes. (Unfortunately, this appealingly neat idea doesn’t really work for Western Oti-Volta, at least as far as those adjectives are concerned which are not obviously derived from dynamic verbs.)

    Thinking more about this specific case, I think my invocation of multiple copulas agreeing with their complements is more than is actually needed to explain the phenomena; it would be more plausible if WOV copulas in general looked liked pronouns, but this is actually something of a stretch, as my examples above probably show.

    My main object was to account for the mismatch between the two languages (Kusaal and Farefare) which have unequivocally verbal copulas; but this could be more parsimoniously accounted for by just positing two original copulas, one for noun complements and one for (non-locative) adverbial complements; the use of any copula at all for adjectives would be a late secondary development.

    Further supporting evidence for this is that even in contemporary Kusaal, rather than an adjective as complement of the copula, a derived adverb (or abstract noun) turns up instead. This would make sense if there was formerly no copula that took an adjective complement. (In fact, pretty much the only adjectives found as complements in Real Life seem to be those which lack derived adverbs and also lack corresponding quality verbs: in Kusaal, this notably includes all the colour adjectives along with “big” and “little”; I don’t know if these are just accidental gaps or if there’s some deep semantic principle at work.)

  101. I asked Michael Weiss about *si; he thinks it was never long in the first place, but that there was an unmarked nom. fem. *si just like the masc. *so (but adding the caveat that “this is speculative beyond my usual speculativity level!”). It’s a strange-looking form, but it does seem to be what the Greek and Gothic data point to.

  102. David Eddyshaw says

    It’s a strange-looking form

    Well, I suppose that there’s no reason a priori why PIE couldn’t have had some forms which were even at that point just plain irregular and unparalleled elsewhere in the declension system. After all, pronouns

    And I suppose *sy is no more peculiar as a word from the PIE-phonology (as opposed to PIE morphology) standpoint than *syH. Who needs vowels anyway?

  103. Trond Engen says

    So what we have is an extension of the useful minor conjugation to nouns by applying it to a pointer? I like that. Would this also work for IE *h1es-?

  104. David Eddyshaw says

    The weirdness of *si might in itself explain why most PIE branches ditched it, I guess.

    There’s also the fact that the whole feminine gender in Indo-European seems to have been a relatively late development, though I can’t work that into the picture very coherently.

  105. Trond Engen says

    Maybe *si existed as a feminine (or polite diminutive?) form before the feminine gender developed. Maybe *h2 was added by analogy. Maybe the existence of a pronoun that lent itself to identification with the *-ih2 pattern was a small but important premise for the development.

  106. David Eddyshaw says

    an extension of the useful minor conjugation to nouns by applying it to a pointer

    I think there’s mileage in that, without straining credulity too far …

    The Farefare copula de looks enough like the singular pronoun of the commonest non-human noun class to be going on with. So that would continue the original copula for nouns.

    The Kusaal copula would then be the one continuing the original copula for adverbs.
    Kusaal aen looks (as I said above) as if it’s based on the root *ŋa, which is a perfectly respectable Proto-WOV plural pronoun; however, there is also a manner-adverb-forming prefix a- (e.g. asida “truly”, from sida “truth”), and that would be the expected outcome for a prefix *ŋa- too.

    Unfortunately, there are, phonologically, two distinct a- prefixes, and although they are identical segmentally, they are preceded by different segmental sandhi (think French h aspiré for an analogy); and this adverb-forming prefix has the wrong sandhi to be derived from *ŋa; it most likely comes instead from *ɲa. I think that the hypothesis could still just about be saved with a bit of special pleading: initial *ɲ seems to have been rather unstable in Proto-WOV, and there are in fact cases where (say) Mampruli has forms derived from initial *ɲ-, whereas Kusaal (and Farefare) have forms going back to *ŋ-, e.g Mampruli nyariŋŋu “boat” versus Kusaal anrʋŋ [ãɾʊŋ], Farefare õorŋɔ; Mampruli, nyariŋŋa “Vitex doniana (tree)”, Kusaal aandig [ã:dɪg], Farefare ãarga. That might then be used to forcibly bring in the Mampruli copula nye as a possible cognate to Kusaal aen after all.

    But it’s all getting pretty epicyclic by that point.

  107. I don’t understand your comment, Trond — the point is that there is no evidence for a *sih₂, only for *si.

    Given its distribution in Western IE (and maybe Indo-Iranian, there’s some relevant data there that I don’t fully understand) it’s probably easiest to assume *si arose at a period when the feminine gender was already established.

  108. Trond Engen says

    You’re right. I got confused and ended up trying to reconcile the wrong bits.

  109. PIE? Polite?

  110. David Eddyshaw says

    Our chariots say we’re polite. You want to dispute it?

  111. Speak laryngeally and carry a big wheel.

  112. David Eddyshaw says

    Mampruli nyasiba “mother’s brother” = Kusaal ansib, Farefare ãsba too …

    The simplest explanation is that it’s Mampruli that has innovated, as all the other WOV languages, even its very close relative Dagbani, agree against it. The evidence from Oti-Volta outside WOV seems to be rather equivocal, but I can’t find many cognates:

    Moba ŋaanŋ “Vitex doniana”‘, Nadwm ɦaadgu, but Waama yɛ̃ribu.
    Moba yiɛ “mother’s brother” (the loss of non-initial *s is regular.)
    (Moba ŋadunŋ “boat” has the wrong tones to be a cognate, and must be a borrowing from WOV, probably actually from Kusaal.)

    More research is needed …

    However, the Dagbani copula is nye, same as Mampruli, and Dagbani does not share this Mampruli ɲ-for-ŋ thing, so unfortunately my idea that the Mampruli copula nye really could be related to Kusaal aen bites the dust. Beautiful hypothesis, meet ugly fact …

    [Hmmm … wonder if it might be conditioned by the original following vowel? (original *ɛ -> a in closed syllables in Mampruli-Dagbani) …]

  113. David Eddyshaw says

    The derivational morpheme could itself be a cliticized copula

    I don’t think so: the WOV minor-conjugation flexion -ya is cognate with the Nawdm flexion -ra, which is found in hundreds of imperfective-only verbs; my favourite cognate is Kusaal zi’e “be standing”, which regularly matches Nawdm jeɦra “be standing” pretty much segment-for-segment (though the usual loss of final vowels in Kusaal makes this less than obvious.)

    So if this was a clitic turned flexion, the process would have to be pre-Proto-Oti-Volta, or at least pre-Proto-WOV-Yom-Nawdm-Buli-Konni (a grouping sadly lacking a proper name as yet, on account of previous incorrect analyses of intra-Oti-Volta relationships: Yom-Nawdm is often wrongly placed as more remote from WOV than the Gurma languages are.)

  114. David Marjanović says

    Further supporting evidence for this is that even in contemporary Kusaal, rather than an adjective as complement of the copula, a derived adverb […] turns up instead.

    WOV substrate in German confirmed.

    And I suppose *sy is no more peculiar as a word from the PIE-phonology (as opposed to PIE morphology) standpoint than *syH. Who needs vowels anyway?

    Perhaps more importantly, it’s no more peculiar than *tw, another personal pronoun that seems beyond dispute (…for Proto-Indo-Tocharian at least; Anatolian points to *ty instead…).

  115. (In fact, pretty much the only adjectives found as complements in Real Life seem to be those which lack derived adverbs and also lack corresponding quality verbs: in Kusaal, this notably includes all the colour adjectives along with “big” and “little”; I don’t know if these are just accidental gaps or if there’s some deep semantic principle at work.)

    “big, little and colour” is a familiar list.

    Cf p. 205 (226)

  116. But colour words and size words differ greatly in this respect in Europe.

  117. David Eddyshaw says

    Yes, it’s tempting to try to find some identifiable reason for these particular adjectives to be different. I think this is probably misguided as far as Kusaal goes, though, because colour adjectives actually do have associated quality verbs elsewhere in Oti-Volta. (On the other hand, if the existing grammars are right – and I see no reason to doubt it – the cases are not quite the same, because in those languages quality verbs are in fact primary, and there aren’t any “real” adjectives, just attributive constructions using those verbs.)

    The “accidental gap” theory is not as far-fetched as you might think, because there are only four colour adjectives in Kusaal (two of which are in fact synonymous, there being three basic colours.)

    Colour adjectives are morphologically distinctive in Arabic, though, now I think of it.

  118. Well, colours and sizes both are adjectives par excellence.

    Colours name physical sensations (and in the link abobe it is big, small, white and insipid).
    Unlike colours, sizes are deeply embedded in language as a system, provide metaphors, change parts of speech and take part in (I do not know how to call this probely) grammatical oppositions like Russian “this shoe to-me small/large” (mal/velik “small/large”) vs. “this shoe is small/large” (malen’kij/bol’shoj “small/large”).

    But both are possibly familiar to shimpanzees.

  119. and little children

  120. Grammatical oppositions – It is semi-regular in Russian. We have two sorts of adjectives in preditative function in present (without copula) and three in other tenses (with copula). In present:
    ya schastliv “I am happy (feel happiness)”
    ya schastlivyj “I am happy (…one, a happy person)”, a bit childish (with this specific word).

    Colour words are rarely used in short versions. Size words, conversely, have developed suppletivism here (compare also “a lot”, “very”, “large”, “big”)

  121. David Eddyshaw says

    I don’t know what motivates the choice of a copula construction as opposed to a quality verb in Kusaal in cases where both constructions exist. In general, quality verbs seem to be a lot commoner when they exist, but I’ve only encountered Li sʋm “It is-good (as opposed to Li an sʋŋa or Li an sʋm) in contexts where it introduces a subordinate clause “It’s good that …”

    … just searched the Bible translation, and that actually does seem to be a genuine rule, rather than a mere tendency; same with Li pʋ sʋm “it’s not good” (there’s no “be bad” verb.) The thing is asymmetrical, though: the form with the copula can introduce a subordinate clause, though it’s more often stand-alone, and it’s not nearly as common for the copula form to introduce a subordinate clause as for the quality-verb construction to do so.

    [it may be something to do with focus. Both sʋm “goodness” and sʋŋa “well” have the peculiar property (shared with a few other words) that they cannot be formally marked as focused (which is why it’s Li an sʋm and not Li anɛ sʋm, in case anybody was wondering); this probably actually means that they are intrinsically focused, which would be inappropriate in most cases when the focus was on a following subordinate clause.]

  122. John Cowan says

    The first sentence of WP s.v. “Grant (law)” is “A grant, in law, is a transfer of property, generally from a person or other entity giving the property (the grantor) to a person or entity receiving the property (the grantee)”, and that confirms the existence of the terms. At the end of the article, we find this: “In legal conveyancing, the grant is the means by which a party conveys title or encumbrance. In trust law, the grant is the act by which the settlor creates the trust for the interests of the trustee. In an option contract, the right of the optionee to exercise the option is considered a grant on the part of the optionor. In philanthropy, a donor may provide a grant of money”, which confirms the generality of the relation. (More authoritative sources back all this up.) Grants for intellectual work aren’t quite philanthropy, but there’s obviously a relation: even if you don’t get your Ph.D., the grantor doesn’t get their money back.

Speak Your Mind