“Look with thine ears”: Why Writing Is Syllable-based

Abstract

Writing, both ancient and modern, originates de novo only when the language written is syllabically organized, because only there does the most salient unit of speech, the syllable, correspond with the most salient unit of language, the morpheme, thus bringing a sound sufficiently to consciousness that a pictograph for the item named by the morpheme can be recognized as also representing its sound – so that it can be transferred to represent the meaning of another morpheme of the same or similar sound (the “rebus principle”).

Why is the syllable the most salient unit of speech? Phoneticians cannot agree on what people recognize when counting or manipulating syllables. Phoneticians can, though, specify what the speech organs do during speech production – the configurations of lips, tongue, larynx, pharynx, and airstream closures that result in the perceived sounds of speech (“consonants” and “vowels”). Phoneticians also recognize “features” whose durations do not necessarily correspond with those of the “segments” that constitute morphemes. Yet it is neither segments nor features that are recorded in earliest writing.

An answer is found in the typology of writing systems, the ways the sounds of language are treated by grammarians and linguists ancient and modern, and the ways devised for children to learn to read.

Keywords:

• origins of writing
• writing systems
• writing system typology
• syllabic basis of writing
• orthographic depth
• phonemic awareness
• stimulus diffusion
• Sumerian writing
• Chinese writing
• Maya writing

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1 I was especially gratified to be invited to participate in the conference “Signs of Writing,” and honored to be, apparently, the first speaker at the first University of Chicago colloquium sponsored by the Neubauer Foundation (8 November 2014).

2 This despite his acknowledgment (276f. n. 39 ad 54) of the work and to some extent favorable reception of the studies by both Whorf and Knorosov; treating them as equivalent suggests less than great care in their study.

3 As Gelb notes (1952, 283f. n. 12), stimulus diffusion was not his term but A. L. Kroeber’s (1940) – see the Excursus below.

4 It is the principles that are important, regardless of whether four subsequent decades of investigation have shown that this or that individual datum may have needed revision. Similar information is found in the more recent articles in a more general journal by David Mora-Marín (2003, 2010), and in the plethora of resources that have since become available, such as Houston (1989), Kurbjuhn (1989), Macri (1996), Macri and Looper (2003), Coe and Van Stone (2005), and Johnson (2013).

5 Cohen shows neither script, instead referring to the illustrations in Diringer’s first edition (1948, 175–177 / 181–184). Diringer interposes the Vai and Mende syllabaries of Africa between the two American ones.

6 The two scripts are not mentioned by Taylor (1883), Clodd (1900), or Mason (1920), but the same pattern is found in the other early surveys considered in Daniels 2002; Faulmann 1880a, 13 / 12; 1880b, 230 / 231; Moorhouse 1953, 121–122 / 122, 163; and in Haarmann 1990, 258–261 / 261–265.

7 The late Suzanne McCarthy suggested that Evans might have been familiar with some South Asian, specifically Tamil, writing (McCarthy 2008).

8 One author, who seemed not to have read anything I’d published outside The World’s Writing Systems (579–586), found this distinction “unhelpful” (Houston 2004, 10), and when pressed for an explanation, replied simply, “I don’t agree with it” (pers. comm. November 9, 2014). The account there (Daniels 1996, 579f., 583, 585) was deliberately kept minimal, as I did not feel a general reference work was a suitable place for setting forth a (then) new theory as if it was authoritative.

9 So limited, in fact, that Reiner (1973) completely disregards both heterograms (for the use of this term, borrowed from Iranistics, in place of : “logogram,” see Daniels 2021; 2018, ch. 7) and semantic determinatives in her algorithm for a hypothetical computer (“programming an information transducer” [8]) to interpret Akkadian cuneiform texts, even though several of each occur in the brief excerpt from Gilgamesh Tablet XI from which she draws her examples.

10 Except, to a limited extent, in an exemplar from Ugarit (Nougayrol 1965) – where, of course, the Ugaritic abjad was in daily use.

11 See spectacular examples in Watson (2002).

12 I am grateful for my title to Zohar Eviatar and John Lithgow. In the Chicago presentation of this talk, the four subtitles were accompanied by Arthur Rackham’s illustrations for Charles and Mary Lamb’s Tales from Shakespeare; identifying the source of each section’s title: 1. The Merchant of Venice (I.iii.20); 2. Macbeth (V.v.21); 3. Romeo and Juliet (I.ii.60f.); 4. King Lear (IV.vi.149).

13 The Excursus was presented at the annual meeting of the North American Association for the History of the Language Sciences, Washington, D.C., January 2016.

14 The first edition cites, in another context, Kenneth L. Pike’s Phonetics of 1943; only in this note is his Phonemics of 1947 added.

15 Such an incidence of deliberate wordplay is rare in Gelb’s extensive corpus.

16 This passage is found in Chapter IX: “We image words, in fact produce them, in syllables, not in sounds. Any one, in slow speech, tends to syllabify, whereas few wholly illiterate people can be induced without patient training to utter the separate consonants and vowels of a word, even for the purpose of teaching a foreigner” (225).

17 Kroeber’s handling of porcelain over the decades is instructive. “China received metals, wheat, cattle and horses, cotton, architecture, religion, possibly the suggestion of script, from the west [sic lowercase]; but she gave to it silk and porcelain, gunpowder and paper” (1923, 426). But subsequently he tips the imbalance even farther: “From western Asia the [Egyptian] art [of glassmaking] was carried to Europe, and, in Christian times, to China, which at first paid gem prices for glass beads, but later was perhaps stimulated by knowledge of the new art into devising porcelain — a pottery vitrified through” (1923, 447f.). “Of course, the distinction of the three great regions of Old World civilization [“China, India, and the Occident”] does not imply that diffusion of culture elements between them ever ceased. … The West, within the historic period, gave glass and perhaps the impulse toward a Chinese ‘invention’ — porcelain, a glazed-through pottery” (467). The Western “impulse” toward the “ ‘invention’ ” of porcelain is absent from the 1940 article. In the second edition of the book the sole mention of glass is in another context; porcelain appears only to exemplify stimulus diffusion; and finally the explanation is given for kaolin, “which on high firing vitrifies all through the ware, instead of only on the surface like a glaze, and hence is brilliant, hard, and waterproof” (1948, 369).

18 Schmidt, using an anthropological approach, identifies three consequences of alphabetolatry that were or would be discussed by reading research scholars as well: the Processing Hypothesis (“the belief that reading processing behaviors are universal for any one script type, independent of social and cultural factors” [4]); the Ease of Learning Hypothesis (“the supposed ease with which alphabets can be learned has been suggested as causal in the development of modern sociocultural systems” [5]); and the Abstraction Hypothesis (“more advanced abstraction skills and metalinguistic insights are needed for the perception of phonemes than syllables in language analysis, skills putatively not available to the ‘primitive’ or non-Western mind” [7]). Schmidt challenges each of them with ethnographic data.

19 Apparently it did not become widely known until the early 1990s that the printed Cherokee characters scarcely reflect the characters Sequoyah designed (Walker and Sarbaugh 1993, 80 fig. 5, 89 n. 7 ad 77, 82–84; see Daniels 2018, 9–11 with fig. 1.1). Much is made of this dichotomy in Cushman (2011).

20 Tuchscherer and Hair (2002) seem very much to want Vai to have been stimulated by Cherokee, but their extensive documentation of the origins of Vai script offers little support for the idea.

21 We now have the advantage of Scribner and Cole (1981) for the script’s present-day uses and Tuchscherer and Hair (2002) for as much as can be known of its origins.

22 Others include grammatical analysis, literary and visual style, and the drama. Quite a few concern external influences on Japan, with a curious footnote enumerating the number of centuries of “lag” of the appearance of various traits in Japan after their origin in China – “block printing, 1; Sung style painting, 2–3; … writing, at least 15; abolition of feudalism, 20” (7 n. 6). This is highly misleading, for Chinese writing was adopted, and not “stimulus-diffused,” for writing Japanese almost as soon as it reached its shores.

23 Gelb (1952, 123) accepts Gardiner’s identification of the West Semitic word bʿlt ‘Lady’ in the Proto-Sinaitic inscriptions but is skeptical about W. F. Albright’s (1948) decipherment, which, though it has been refined over the decades, laid the groundwork for their interpretation.

24 Kroeber provides a six-page appendix detailing the correspondences between the two editions (851–856).

25 Kroeber notes (1948) that the chapter “Story of the Alphabet” is a “revision” of “The Spread of the Alphabet,” “which was largely based on I. Taylor [1883]. The only new section is 207, on Anticipations of the Invention [advancing the extraordinary hypothesis that Ugaritic writing could have been an independent invention of the consonantal (abjadic) principle – aided by Z. S. Harris’s (e.g. 1939) acceptance of the initial dating of the texts several centuries too early]. Changes mainly follow H. Jensen, Geschichte der Schrift, 1925.” Among the changes to the quite beautiful map of the spread of alphabetic writing (1923, fig. 30 facing p. 284 [reproduced in Daniels 2018, 190 fig. M.1] >1948, 530 fig. 26) is the addition of a connection from the Greek alphabet to the Armenian – but not the deletion of the line from Pehlevi to Armenian (cf. the note to Figure 2 on p. 95 above). Published in the same year, Diringer (1948) discusses Armenian in the same chapter as Pehlevi, as does the later edition Diringer (1968), though with some skepticism as to Junker (1925–26), which may have been regarded as the definitive last word in the first edition. Kroeber’s “revisions,” however, go beyond mere factual matters that might have come from Jensen, and include the addition of a brief introductory passage noting that “the history of the alphabet thus illustrates at one point or another the principles of basic invention, supplementary invention, diffusion, acceptance, refusal, modification, survival, loss, patterning, and function — and the interweaving of all these” (509): that is, the principles Kroeber saw as operating throughout human culture. This points up the inadequacy of treatments of writing through most of the 20th century: his earlier authority was 40 years old; his later one, nearly 25 at the time he was writing.

26 The second edition shows inconsequential changes acknowledging the decipherment of Linear B, which had intervened since the first edition.

27 This is a strictly geographic name, like “Proto-Sinaitic.” There is no reason to assume that “Proto-Elamite,” which Gelb dates 3000–2200 bce, records a language akin to the Elamite known from cuneiform inscriptions preserved from two disparate periods, the 13th c. bce and the 7th c. through the Achaemenid era that came to an end with Alexander’s conquest (Reiner 1969; Stolper 2004), a point not made by Gelb.

28 I.e. Harappan or Indus Valley.

29 I.e. Linear B and several other not then and not yet deciphered scripts, as well as those from Cyprus.

30 I.e. Luvian. As early as his first essay on the decipherment of the “Hittite hieroglyphs,” Gelb (1931, 82) recognized that “two languages, Luvi and Palai, … might be regarded as possibly identical with that of the hieroglyphic inscriptions,” but as evidence mounted connecting them with the former, true to form he never changed the label he used for the writing.

31 Puett advocates a middle ground that recognizes indigenous developments as well as imported features (metallurgy, chariot warfare, cavalry formations); he does not include writing among imported features.

32 Citing Terrien de Lacouperie from 1887–94 and C. J. Ball from 1913, when almost nothing was known of the earliest attested stages of either writing.

33 The literature cited by Gelb for this point belongs equally to the period identified by Puett (1998) as the heyday of the search for external influence on China.

34 The unclarities of Gelb’s description of Egyptian writing (72–81) are set forth in Daniels (2000).