Script origin and typology, part 1

« previous post | next post »

[This is a guest post by Peter T. Daniels]

Author's Note

In 1999, Holly Pittman of the University of Pennsylvania invited me to prepare a talk to close an international symposium on early writing systems. The result is before you — essentially unchanged and unupdated (because the planned publication did not materialize), even though I would treat a couple of points differently now. John Noble Wilford covered the event for the New York Times, but in order to accommodate illustrations, his article was cut (from the bottom, as newspapers do), and since he described each contribution in the order it was given, the last several talks went unmentioned! (And weren't restored when a volume of his reporting was published a few years later.) 

A fuller presentation of my understanding of the nature and history of writing may be found in my Exploration of Writing (Equinox, 2018), and in major articles in the 2023 volumes of the journals WORD and Written Language and Literacy.

A Study of Origins
Peter T. Daniels
New York [now Jersey City, N.J.]

closing talk at The Multiple Origins of Writing: Image, Symbol, Script
international symposium, Center for Ancient Studies,
University of Pennsylvania. University Museum, Philadelphia, March 27, 1999

On February 11, 1859, the following words were delivered by a man who was on the verge of worldwide acclaim:

Writing—the art of communicating thoughts to the mind, through the eye—is the great invention of the world. Great in the astonishing range of analysis and combination which necessarily underlies the most crude and general conception of it—great, very great in enabling us to converse with the dead, the absent, and the unborn, at all distances of time and of space; and great, not only in its direct benefits, but greatest help, to all other inventions.

. . .            When we remember that words are sounds merely, we shall conclude that the idea of representing those sounds by marks, so that whoever should at any time after see the marks, would understand what sounds they meant, was a bold and ingenious conception, not likely to occur to one man of a million, in the run of a thousand years. And, when it did occur, a distinct mark for each word, giving twenty thousand different marks first to be learned, and afterwards remembered, would follow as the second thought, and would present such a difficulty as would lead to the conclusion that the whole thing was impracticable. But the necessity still would exist; and we may readily suppose that the idea was conceived, and lost, and reproduced, and dropped, and taken up again and again, until at last, the thought of dividing sounds into parts, and making a mark, not to represent a whole sound, but only a part of one, and then of combining these marks, not very many in number, upon the principles of permutation, so as to represent any and all of the whole twenty thousand words, and even any additional number was somehow conceived and pushed into practice. This was the invention of phoenetic writing, as distinguished from the clumsy picture writing of some of the nations. That it was difficult of conception and execution, is apparent, as well by the foregoing reflections, as by the fact that so many tribes of men have come down from Adam’s time to ours without ever having possessed it.

Charles Darwin, on the eve of publication of The Origin of Species? No: Abraham Lincoln, on the day before his—and Darwin’s—fiftieth birthday. In this passage from a “Lecture on Discoveries and Inventions,” given at Illinois College in Jacksonville, and then Decatur and Springfield, the failed politician and future president anticipated what I have to say about the origins of writing.

The book universally recognized as the founding document in the scientific investigation of writing systems is I. J. Gelb’s A Study of Writing. According to the Preface, much of it was written in the late 1930s, thus making it contemporary with Joseph Vachek’s pioneering work on written language. My own work on the typology and origins of writing is very much a reaction to various claims in that book; unfortunately Gelb did not live to see and comment on my suggestions. But my approach also resembles that of another contemporary project: Arnold Toynbee’s A Study of History. Toynbee took the whole world and all its history as his subject matter, and he believed he found trends and parallels in the development of all the civilizations he identified (twenty-one of them, in his final inventory). Invoking this grandiose achievement is not in fact so far-fetched as might be supposed, for among the dozens of topics to which individual essays are dedicated throughout the ten volumes and two supplements are phenomena of, in order, “Lingue franche,” “Archaism in Language and Literature,” and “Official Languages and Scripts.” Even in light of Toynbee’s university training as a Classicist, these essays are impressively acute, especially considering their dates of publication of 1939 for the first two and 1954 for the third. Toynbee evidences a special interest in the Achaemenid period generally, and things Iranian have recently proved to be particularly illuminating of script-related questions overall. What I have done, Toynbee-like, is to observe the nature and behavior of the full range of writing systems around the world and, Gelb-like, extract a typology and enunciate general principles.

It all started with dissatisfaction with Gelb’s claim that the Phoenician and related scripts are not alphabets, but “syllabaries with unspecified vowel”; and that Ethiopic and Indic scripts are not syllabaries, but some form of alphabet. Eventually I realized that these claims had to be made not because of some inherent properties of the scripts in question, but in order to save Gelb’s “Principle of Unidirectional Development”—which claims, without justification and, in fact, contrary to fact, that the three types of writing system, logosyllabic, syllabic, and alphabetic, must succeed each other in that order and without exception. But if Cypriote or Linear B or the Japanese kana are syllabaries, then Phoenician is not a syllabary. If Greek and all its descendants and Korean are alphabets, then Ethiopic is not an alphabet. The Principle of Unidirectional Development simply does not hold.

But if we recognize that there are two different kinds of writing systems that denote complete syllables, half the problem goes away. There are the true syllabaries, like the Linear B, the Cypriote, or the kana, with a distinct character for each possible combination of a consonant plus a vowel (maybe 50 to 80 different signs). Then there are the other kind, the Ethiopic and the Indic family, with relatively a smaller number of distinct characters, and each one has a basic form, and the basic form denotes a consonant plus the unmarked vowel (generally /a/), and each of the other vowels is denoted by the addition of some particular mark or modification to the basic form. A variety of names for this type of script has appeared in the literature—alphasyllabary, neosyllabary, pseudoalphabet—but I reject any name that incorporates either “alphabet” or “syllabary,” since they suggest subtype-ness or dependency, and I wish to stress independence. The name I use for this type is abugida, an existing Ethiopic word combining the first four consonants of the traditional Semitic order with the first four vowels of the traditional Ethiopic chart. With this distinction of two different kinds of syllable-denoting writing systems, half the problem of Gelb’s Principle goes away.

However, the Phoenician script is still neither a true syllabary nor an abugida. But it is also not a full alphabet, because of the want of signs for vowels. It constitutes another separate type of script, a consonantary, and my name for this type is abjad. Abjad is an existing Arabic word for the letters of the Arabic script taken in the traditional Semitic order; I use these words abugida and abjad because of their structural resemblance to the word “alphabet.” In ordinary contexts, both Ethiopic and Arabic use other standard orders of the script, but the traditional Semitic order is known in Ethiopic because letter names appear in the Bible over the sections of Psalm 119, and in Arabic because the numerical values of the letters reflect the historic order rather than the rearrangement according to their shape. Now with a set of five types of writing system (plus a sixth, the featural, that comes into play in various derived scripts) I can describe the historical development of scripts with a scheme that is less neat, but more accurate, than Gelb’s tripartite Principle of Unidirectional Development.

For the next step, I turn from a Gelb-like activity to a Toynbee-like activity. Since I’m ultimately interested in the original origins of writing, I look at as many modern inventions of writing as I can find. Descriptions of most of them are collected by Alfred Schmitt, in the posthumously published Entstehung und Entwicklung von Schriften (1980). At first it looks as though any kind of writing can get invented anywhere. Cherokee of the southeast United States is a syllabary; Cree of northern Canada is a featural abugida; Vai of west Africa is a syllabary; Tolkien’s Tengwar is a featural alphabet. But if we consider the inventors of the various scripts about whom we have some information, it turns out these are two quite different sorts of people. Some of them could already read in some language and even had at least some education in phonetics. Such people include James Evans, responsible for the Cree script; King Sejong and his advisors, who created Korean Hangul; and, of course, J. R. R. Tolkien, whose entire fantasy world was devised to make possible play with invented lang­uages.

But much more interesting in investigating the original origins of writing are those inventors who knew nothing about writing, except the fact that it existed: that a piece of paper could, apparently, talk to someone without a voice. Such were Sequoyah, inventor of Cherokee script; Mɔmɔlu Duwalu Bukɛlɛ, inventor of Vai script; Uyaqoq of Alaska; Afaka of Suriname; and a number of others: each of these was, as Lincoln put it, “one man of a million, in the run of a thousand years.” Clearly these untutored writing inventors—or as I say, grammatogenists—are closer to the ancient grammatogenists, creating before the discovery of grammar. And do all their scripts—Cherokee, Vai, Alaska script, Njoka, Barnum, Caroline Islands, and the others—do they have anything in common? Yes, they do: They are all syllabaries, Consonant–Vowel syllabaries. They comprise about a hundred distinct characters, each denoting what was still in the time of Lincoln called a sound—a complete syllable. Such sounds are the basic units of speech. They are what people naturally break words into if they haven’t been taught by their alphabet to find smaller segments—what Lincoln called parts of sounds.

And, such sounds are what the first writers first wrote down. Until recently, we could only read two ancient scripts that could fairly securely be believed not to share an origin: Sumerian cuneiform and Chinese oracle bone. Ever since we’ve been able to read those two scripts, we have known they are both logographic—a sign represents a word. Or, as Gelb taught us to say, they are logosyllabic—a sign represents both a word and a syllable. Or, more precisely, they are morphosyllabic—a sign represents both a morpheme and a syllable. Until recently this could be seen as mere coincidence: the two earliest scripts were devised for languages where each morpheme is just one syllable. But then came the decipherment of Maya writing, and we knew of a third original script for which the same is true. In its “inner form,” Maya writing is very like Sumerian, with a logographic core to which affixes are appended—phonetic rather than grammatical. All three scripts are syllabic. They are not segmental, and they are not abugidic.

Modem grammatogeny is not the only evidence that syllables are something special in the stream of speech. Developmental psycholinguists, and scholars of literacy and its acquisition, have conducted experiments on literates and nonliterates and preliterates and illiterates, and their findings all seem to converge on the basicness of the syllable (as opposed to the segment). Instrumental phonetics shows that nothing in the stream of speech corresponds to the segments written with alphabets. The syllable is clearly the most salient minimal stretch of speech. Similarly, as Lincoln also knew, the word is the most salient minimal stretch of language.

So now we put these observations together. Untutored grammatogenists, ancient or modern, create scripts that record syllables. Modern creators did so by breaking words into their smallest sounds: their syllables. For ancient creators, just as Lincoln supposed, “a distinct mark for each word, … would follow as the second thought”; but, as Lincoln could not know, in some languages, the mark for each word would also be a mark for an individual sound (a syllable), and so would not “present such a difficulty as would lead to the conclusion that the whole thing was impracticable.” That is, in some languages, the morphemes are generally monosyllabic; meaning that putting down a mark for a word is also putting down a mark for a syllable; and that same syllable might also be the sound of a different word, or very like it.

Herein lies the key to inventing a writing system. It is not unusual for non­literate peoples to make graphic representations of meaning that are, however, not writing, because they do not represent specific language. Pictographs—Lincoln’s “clumsy picture writing of some of the nations”—are found around the world; but we do not “decipher” them, for we recognize that they do not convey individual words: they may correspond to individual words, but they stand for things or events. But in a monosyllabic language, a picture corresponding to a word, standing for a thing, also stands for the sound of that word; and it might also stand for the sound of a different word, the same or a very like sound; and this different word might be the word for something it’s not so easy to draw a picture of. A standard example, due to Gelb, is that in Sumerian a picture of an ‘arrow’, pronounced ti, could also be used to represent ‘life’, also pronounced ti. This, of course, is the rebus principle that underlies logosyllabic writing.

It is the rebus principle that makes possible true, full writing—since writing is a system of more or less permanent marks used to represent an utterance in such a way that the utterance can be recovered more or less exactly without the intervention of the utterer. This definition means that writing must be able to convey everything in language that is not concretely picturable: the grammatical morphemes, the names (especially foreign ones) that have no secular meaning. It means that some at least of the characters in the script must be used sometimes for their sound values alone: at the earliest stages of true writing, there must be a pure syllabic component to the script. Sumerian with its monosyllabic morphemes was well suited to lend at least some of its signs to this syllabary.

If, however, a language is not monosyllabic—as in, for instance, Indo-European or Semitic or Uralic or Altaic—the chances are rather less good that the picture put for one word would have the same sound as another word or one very like it, as with the Sumerian ti example. And that is why writing could get started in Sumerian, in Chinese, in Maya, and probably in Dravidian; while the best candidate for writing where it didn’t get started—the Inca civilization—did not use a monosyllabic language, and so came up with quipus for accounting, but not with writing. Maybe there were pretty complex cultures all over the world, and maybe they routinely drew pictographic aides-memoire, but if the languages weren’t monosyllabic, it was too big a leap to make the signs for some words represent other words according to their sounds.

[to be continued]

 

Bibliography

[Will be repeated at the end of part 2.]

Bühler, Georg. 1898. On the Origin of the Indian Brāhma Alphabet, together with Two Appendices on the Origin of the Kharoṣṭhī Alphabet and the Origin of the So-Called Letter-Numerals of the Brāhmī. 2nd ed. Strassbourg: Trübner.

Daniels, Peter T. 1990. “Fundamentals of Grammatology.” Journal of the American Oriental Society 110: 727–31.

Daniels, Peter T. 1992. “The Syllabic Origin of Writing and the Segmental Origin of the Alphabet.” In The Linguistics of Literacy, edited by Pamela Downing, Susan D. Lima, and Michael Noonan, 83–110. Typological Studies in Language 21. Amsterdam and Philadelphia: John Benjamins.

Daniels, Peter T. 1999. “Some Semitic Phonological Considerations on the Sibilants of the Greek Alphabet.” Written Language and Literacy 2(1): 57–61.

Franklin, Benjamin. 1987. Writings, edited by J. A. Leo Lemay. Library of America 37. New York: Literary Classics of the United States.

Gelb, I. J. 1952. A Study of Writing: The Foundations of Grammatology. Chicago: University of Chicago Press. (2nd ed., 1963.)

Henning, W. B. 1958. “Mitteliranisch.” In Iranistik, 20–130. Handbuch der Orientalistik I/4.1. Leiden: Brill. (Unpub. English translation by Peter T. Daniels available.)

Kara, György. 1996. “Aramaic Scripts for Altaic Languages.” In The World’s Writing Systems, edited by Peter T. Daniels and William Bright, 536-58. New York: Oxford University Press.

Lincoln, Abraham. 1859. “Lecture on Discoveries and Inventions, Jacksonville, Illinois.” In Abraham Lincoln: Speeches and Writings, 2 vols., edited by Don E. Fehrenbacher, 2, 3–11. Library of America 45–46. New York: Literary Classics of the United States, 1989.

Sampson, Geoffrey. 1985. Writing Systems. London; Stanford: Hutchinson; Stanford University Press. (Corrected pbk. reprint, London, 1987.)

Schmandt-Besserat, Denise. 1992. Before Writing. 2 vols. Austin: University of Texas Press.

Schmitt, Alfred. 1954. “Die Vokallosigkeit der ägyptischen und semitischen Schrift.” Indogermanische Forschungen 61: 216–27.

Schmitt, Alfred. 1980. Entstehung und Entwicklung von Schriften, edited by Claus Haeber. Cologne: Böhlau.

Skjærvø, P. Oktor. 1996. “Aramaic Scripts for Iranian Languages.” In The World’s Writing Systems, edited by Peter T. Daniels and William Bright, 515–35. New York: Oxford University Press.

Stuart, George E. 1992. “Quest for Decipherment: A Historical and Biographical Survey of Maya Hieroglyphic Investigation.” In New Theories on the Ancient Maya, edited by Elin C. Danien and Robert J. Sharer, 1–63. University Museum Monograph 77, University Museum Symposium Series 3. Philadelphia: The University Museum, University of Pennsylvania.

Toynbee, Arnold J. 1939a. “Lingue franche.” In A Study of History, pt. V: The Disintegrations of Civilizations, sec. C: The Process of the Disintegrations of Civilizations, 1: The Criterion of Disintegration, d: Schism in the Soul, 6: The Sense of Promiscuity, γ, vol. 5, 483–527. London: Oxford University Press.

Toynbee, Arnold J. 1939b. “Archaism in Language and Literature.” In A Study of History, pt. V: The Disintegrations of Civilizations, sec. C: The Process of the Disintegrations of Civilizations, 1: The Criterion of Disintegration, d: Schism in the Soul, 8: Archaism, γ, vol. 6, 62–83. London: Oxford University Press.

Toynbee, Arnold J. 1954a. “The Administrative Geography of the Achaemenian Empire.” In A Study of History, pt. VI: Universal States, sec. C: Universal States as Means, II: Services and Beneficiaries, c: The Serviceability of Imperial Installations, 3: Provinces, Annex, vol. 7[B], 580–689. London: Oxford University Press.

Toynbee, Arnold J. 1954b. “Official Languages and Scripts.” In A Study of History, pt. VI: Universal States, sec. C: Universal States as Means, II: Services and Beneficiaries, d: The Serviceability of Imperial Currencies, 1, vol. 7[A], 239–55. London: Oxford University Press.

Vachek, Josef. 1939. “Zum Problem der geschriebenen Sprache.” Travaux du Cercle Linguistique de Prague 8: 98–104. (Repr. in his Prague School Reader in Linguistics [Bloomington: Indiana University Press, 1964], 441–52; rev. Eng. trans. in his Written Language Revisited [Amsterdam and Philadelphia: John Benjamins, 1989], 103–16.)

Wachter, Rudolf. 1991. “Abbreviated Writing.” Kadmos 30(1): 49–80.



19 Comments »

  1. Philip Taylor said,

    July 2, 2024 @ 3:12 am

    Excellent. I look forward very much to part 2.

  2. KIRINPUTRA said,

    July 2, 2024 @ 6:23 am

    +1

  3. Jonathan Smith said,

    July 2, 2024 @ 9:42 am

    One reflection on this (increasingly standard?) terminology is that syllabic and logosyllabic systems aren't related in the way the terms might suggest. That is, so-called "logosyllabic" systems are emphatically not "syllabic" in the technical sense — they're fundamentally logographic, with the addition of "-syllabic" just pointing up a (descriptively if certainly not historically) incidental feature of the words at issue viz. they are (often) monosyllabic. Given this difficulty I find myself saying things like "syllables of sound", "sound-syllabic", etc., to be clear about the nature of true syllabic i.e. kana-type systems. Alternatively or simultaneously we could make up e.g. "1σ-logographic" to be clearer about the nature of the mapped elements in cases like Chinese (= NOT phonological syllables).

    Re: Chinese in particular, this "incidental" feature certainly seems to have been crucial to the system's expansion prior to the OBI period — but the exact role of monosyllabicity is maybe not so clear. Whatever one thinks about the idea of an early Heavenly Stems + Earthly Branches logography per se, it seems probable that one or more proto-writing schemes of this general kind constituted the germ(s) of the larger script and that the individual dozen or so words implicated in such systems need not have been monosyllabic (yet). From such a POV, it looks reasonable that rebus and similar devices are not exploited within the Stems characters but do appear e.g. within the Branches and the numerals as coverage grows.

  4. Scott P. said,

    July 2, 2024 @ 3:25 pm

    It seems odd to this layperson to describe Sumerian as 'monosyllabic' when it is an agglutinative language. To take one example, in the tablet of the discussion between Enki and Ninhursag, the line, "there was a city — the one we live in" is transliterated "iriki na-nam na-an-dur2-ru-ne-en-de3-en" — that's three words and 13 syllables.

  5. Cirk R. Bejnar said,

    July 2, 2024 @ 4:36 pm

    Very good overall. In particular I have seen the abjad and abugida language used elsewhere as standard in ways which are clear and helpful. The discussion of Maya writing strikes me as as slightly incomplete and possibly misleading.

    Firstly, the Classic Maya script was probably derived in large part from earlier systems (e.g. Epi-Olmec/Isthmian) that recorded a non-Maya language. Though this hasn't been demonstrated it is at least possible that Maya is more like Akkadian than Sumerian cuneaform in regards to its relationship to the invention of writing.

    Secondly, Maya is not monosylabic in the sense used in the paper. Most words have a CVC or even CVCVC structure that can make phonetic spellings ambiguous even without considering any grammatical affixes that may be present.

  6. Chris Button said,

    July 4, 2024 @ 2:03 am

    @ Jonathan Smith

    Could you possibly explain what you mean a little more when you say this:

    it seems probable that one or more proto-writing schemes of this general kind constituted the germ(s) of the larger script and that the individual dozen or so words implicated in such systems need not have been monosyllabic (yet).

  7. Chris Button said,

    July 4, 2024 @ 2:16 am

    If, however, a language is not monosyllabic—as in, for instance, Indo-European or Semitic or Uralic or Altaic—the chances are rather less good that the picture put for one word would have the same sound as another word or one very like it, as with the Sumerian ti example. And that is why writing could get started in Sumerian, in Chinese, in Maya, and probably in Dravidian; while the best candidate for writing where it didn’t get started—the Inca civilization—did not use a monosyllabic language, and so came up with quipus for accounting, but not with writing.

    I'm not sure I fully buy this–particularly if "monosyllabic root" is what is meant.

    (Incidentally, aren't there interesting theories that quipu perhaps served for more than just accounting? Albeit not ever being a writing system of course)

  8. Jonathan Smith said,

    July 5, 2024 @ 12:13 pm

    @ Chris Button

    I mean that if Chinese writing developed towards the OBI stage over many centuries beginning from one or more small symbol sets having emerged in relation to very specific functions (e.g., to write the ten 'Stems' calendrical terms), I don't see monosyllabicity as a necessary property of such terms in order for the associated logographs to have emerged… indeed I don't see why we should impose any particular expectations re: word form for this stage as these are just (hypothetical) bags of logographs among which "rebus"-type devices have yet to be exploited.

  9. Chris Button said,

    July 6, 2024 @ 3:41 am

    I think connections with tangible concepts/things (e.g. 卯 with 劉, 巳 with 子, 戌 with 戉, etc.) would probably render that unlikely.

  10. Jonathan Smith said,

    July 6, 2024 @ 7:08 am

    ?? I pointed out that the TGDZ were regular old words of early Chinese back in 2010, if that's what you mean… an idea which strengthens earlier, more general proposals that these systems played a role in the development of "proto-writing" in China.

  11. Chris Button said,

    July 6, 2024 @ 9:58 am

    I'm not suggesting any etymological connection. The ganzhi have no attested meanings.

    Rather, I am saying that there are connections in shape and sound.

    As to why they were selected from the script for that purpose remains obscure. Was there external influence in that regard–particularly in terms of their representation of onsets (https://languagelog.ldc.upenn.edu/nll/?p=64822#comment-1619861) ? How were they chosen in Egyptian for example? Is that known?

  12. Zev Handel said,

    July 10, 2024 @ 5:32 pm

    Re: Jonathan Smith's comments on the monosyllabicity (or lack thereof) of early Chinese writing: Recent trends in Old Chinese reconstruction, if correct, point to a very different analysis of the linguistic representational value of Chinese characters. The word structure proposed for Old Chinese by Baxter & Sagart (among others) involves a syllabic (CVC) root with potentially four (or even more?) morphological affixes. These affixes (prefixes, suffixes, infixes) might be sub-syllabic consonants, so that the resulting derived word is one-syllable long, or in the case of prefixes might be light syllables, so that the resulting word becomes longer than a syllable. Yet no matter how morphologically complex or how many syllables the word becomes, it is still written with a single character.

    I'll just give two examples. B&S reconstruct the Old Chinese word for 'to feed' as *s-m-lək-s (a monosyllabic root with two prefixes and one suffix). It is written with the single Chinese character 飼. B&S reconstruct the Old Chinese word for 'shade' as *mə-q[u]m-s (a monosyllabic root with one prefix, one infix, and one suffix). The prefix is a light syllable, so that the whole word is two syllables long (or, if you prefer, sesquisyllabic). It too is written with the single Chinese character 蔭.

    Moreover, Baxter & Sagart's system allows for the possibility of variant pronunciations so that 'to feed' could be sometimes pronounced *sə-mə-lək-s. Three syllables.

    This makes early Chinese writing look much more logographic, in the etymological sense of the term. That is, individual characters are writing complex words rather than single morphemes. It also makes it difficult to characterize the writing as logosyllabic, since many characters write words that are longer than one syllable.

    One could still argue that there is a level of abstraction operating within the writing system, at which each character represent a monosyllabic root morpheme, and this level determines how characters are repurposed via the rebus principle. But in actual practice, the characters are frequently representing words of morphological complexity and polysyllabic weight.

    How to characterize such a writing system? I'm not sure.

    Oddly, it is later Chinese writing (from about the Han dynasty on) that seems to more closely resemble the morphosyllabic form that is presumed to be typical for ex nihilo inventions of writing.

    Now, one could turn things around and say, What we know about early writing tells us that Chinese writing should have started out as morphosyllabic. So Baxter & Sagart's speculative reconstructions must be wrong. But that's not a very solid foundation for an argument. There aren't enough examples of invented writing to establish definitive principles, and there is a lot of independent evidence for B&S's reconstruction system.

  13. Pamela said,

    July 10, 2024 @ 7:26 pm

    To my completely ignorant eye, this description of multi-syllabic ancient Chinese seems very like Tibetan (presumably ancient Tibetan). Is that a connection? What is the chronology like? When is "early Chinese" –does it follow proto-Sino-Tibetan?

  14. Guillaume Jacques said,

    July 11, 2024 @ 4:26 am

    Following up on Jonathan and Zev's discussion, I think that the only strong argument for the equivalence one character = one syllable for pre-Zhanguo periods is the regularity of number of characters in the verses of the Shijing. In this regard, it is interesting to listen to how languages with sesquisyllables actually treat them in songs. I remember listen to David Sangdong (a Rawang scholar, author of this dissertation: https://opal.latrobe.edu.au/articles/thesis/A_grammar_of_the_Kadu_Asak_language/21847302) singing in Rawang, and was baffled to see that syllables with presyllables just counted as one syllable (the presyllable simply did not count in the meter). This offers a good model of how even with presyllables, Old Chinese may have a regular number of characters in verses. In addition, there is the problem that verses in bronze inscriptions have a notoriously irregular number of syllables per verse (see Wolfgang Behr's work), something that has not be satisfactorily explained.

  15. William Hannas said,

    July 11, 2024 @ 6:08 am

    Second Pamela's comment about the similarity of these reconstructions to the morphology of pre-modern Tibetan as captured in today's (highly conservative) orthography, which crowds consonant prefixes, infixes, and suffixes around nuclear vowels, the whole of it expressed (by small dots at the end) as monosyllables Many of the consonants are not pronounced (or get reduced) in modern Tibetan speech. Were they ever pronounced? If so, it's hard to imagine that happening without additional vowel support of some kind. Apologies as it's been decades since I looked at "Sino-Tibetan" scholarship.

  16. Chris Button said,

    July 11, 2024 @ 6:12 am

    Regarding sesquisyllables/presyllables, my system for Old Chinese only acknowledges them in fusions or (pre-)OC loanwords. A couple of examples, both discussed elsewhere on LLog:

    車 ᵏɬaɣ from earlier kə̯ᵏlaɣ (as a loan from Tocharian)

    七 ʦʰǝjs from earlier ᵏrʰə̯ʣǝjs (as a fusion of 六 rəkʷ ~ ᵏrʰəkʷ with 二 ⁿʣǝjs as attested in Old Burmese "second six")

    Personally I'm not partial to the extensive affixation proposed by Baxter & Sagart.

  17. Chris Button said,

    July 11, 2024 @ 2:28 pm

    @ Zev Handel

    But in actual practice, the characters are frequently representing words of morphological complexity and polysyllabic weight.

    How to characterize such a writing system? I'm not sure.

    Perhaps the problem is with the nature of the reconstruction rather than with the classification of writing systems?

  18. Jonathan Smith said,

    July 11, 2024 @ 3:59 pm

    More speculation in light of comments…

    * Even in versions of "OC" which feature lots of disyllabic (or longer!) lexemes, much of the lexicon remains monosyllabic and these forms serve in effect as template for tonic syllables ≈ roots of longer words — so re: the question of the relationship between typology and the development of writing, one could IMO still argue with Daniels that this state of affairs provided fertile ground for the expansion of writing in the form of lots of relatively readily exploitable homophonies and near-homophonies at a syllable≈word-ish level. This is the kind of situation I had in mind in my first post as perhaps being "crucial to the system's expansion" pre-OBI… not exactly a testable hypothesis though :(

    * If your favored OC as of say pre-OBI has maximally /CCVC/, that is, any "sesquisyllabic" word forms are just phonetic realizations of particular onset clusters, then you don't need even the above argument — you just have monosyllables and Daniels' argument such as it is is in good shape if no more testable.

    * Re: Shijing, which I've mentioned somewhere(s) without really knowing how to interpret it, Guillaume's reference to Rawang etc. verse is of much interest. Perhaps we can, perversely, define syllabicity (in part) with respect to verse here, concluding that all words are phonological monosyllables despite what our lying ears may think about phonetics.

  19. David Prager Branner said,

    July 11, 2024 @ 4:17 pm

    I agree with @Zev Handel, that reconstructed forms like Baxter-Sagart's *mə-q‹r›[u]m-s make early Chinese look more logographic than Schuessler's *ʔəms does. *mə-q‹r›[u]m-s is unique in their system, while *ʔəms not only represents 蔭 "shade: where trees or hills block the sun" but also 喑 "silent: unable to produce sound", and 闇 "dark: lacking light", and the transitive sense of 飲 "to water: to give drink to [an animal or person]".

    Two things are worth remembering about the Baxter-Sagart forms. One is that they are neither reconstructions nor transcriptions in the senses of those words usual in connection with other languages. Why? Because they are not derived from actual, spoken words of any era nor from phonologically transparent written representations of actual words. Rather, they are highly abstract packages, studded with etymological information placed in them by the authors of the system. They are more like fruitcakes, full of nuts and candied citron peel, and possibly soaked in rum or bourbon, than they are like pieces of fruit picked from a tree.

    The other thing to remember is that morphological affixation is not a conclusion of modern linguistic sinology. No, it has been a premise of the field, from before the era of reconstructions. Already in 1894, well before the pioneering work of Karlgren, Otto Jespersen had proposed that Chinese tonal doublets (such as the twin readings of 陰, one of which is also written 蔭) must have originated in now-lost "derivative syllables or flexional endings and the like". And from the 1860s onward, if not before, Western scholars such as Karl Lepsius and Wilhelm Grube had been arguing that Chinese monosyllabism must have been the end result of a process of reduction of more complex words, of more than one syllable. Morphological affixation had already been in the air, surrounding Western study of Chinese, well before modern reconstructions began being published.

    So I have reservations about using Baxter-Sagart to draw conclusions about what Chinese actually looked like at an early era. Their forms are interesting as vessels for speculative ideas, but not as the representation of real words. One of those speculations is that fully ramified Chinese syllables were more likely to be unique (and therefore more logographic) than in later periods. On reflection, I think that's also the point @Chris Button is making.

RSS feed for comments on this post

Leave a Comment