Whence cometh linguistic meaning?

« previous post | next post »

Sino-Platonic Papers is pleased to announce the publication of its two-hundred-and-eighty-seventh issue:

“Emotion, Reason, and Language: Meanings Are Made, Not Retrieved” by J. Marshall Unger.  A free pdf of this paper is available here.


Most researchers of writing are no doubt familiar with the historical, general linguistic, and neurolinguistic arguments in works by DeFrancis (e.g. 1984, 1989), Unger (e.g. 2003, 2014, 2016), and others against the conceptualization of writing systems as either phonographic or logographic, and thus against the disjoint classification of their graphic units as either phonograms or logograms. Here I discuss new studies of reasoning and conscious thought (especially Barrett 2017 and Mercier & Sperber 2017) that shed light on the classification of writing systems and their components from yet another direction. These studies make new and cogent arguments that emotion and reason are not stored routines or algorithms retrieved by mental processes akin to computer programs but are rather thoughts and behaviors that emerge in real time. Recent studies of brain states during the transition from word recognition to word understanding (Kutas & Federmeier 2011, Broderick et al. 2018) strongly suggest that linguistic meaning likewise is dynamically created on the occasion of each linguistic interaction. Given the failure of the metaphorical understanding of emotions, reasoning, and lexical meanings as prestored context-independent entities retrieved during cognitive activity, the notion of a pure logogram — a symbol representing a unit of meaning possessed by an individual word or morpheme — becomes vacuous.

On the phonographic-logographic spectrum of the conceptualization of writing systems, I must confess that I am on the phonographic side, definitely not the logographic side, whereas Unger tries to bridge the gulf between the two sides.


Selected readings

"The horror of ideograms" (2/25/09)

"Pinyin for the Prez" (10/25/18)

"'Arrival is a tree that is still to come'" (10/10/16)

"Firestorm over Chinese characters" (5/23/16)

"Which is worse?" (1/21/16)

"Chineasy? Not" (3/19/14)

"Writing: from complex symbols to abstract squiggles" (6/11/19)

"Homophonophobia" (2/7/15)

"Zhou Youguang 1906-2017" (1/14/17)

"John DeFrancis, August 31, 1911-January 2, 2009" (1/26/09)

"Miracle" (12/21/14) — especially in the comments

"Dumpling ingredients and character amnesia" (10/18/14)


  1. Antonio L. Banderas said,

    June 14, 2019 @ 11:25 am

    Could somebody please elaborate a bit on the "phonographic-logographic spectrum of the conceptualization of writing systems"? I'd appreciate some examples.

  2. Antonio L. Banderas said,

    June 14, 2019 @ 11:43 am

    Secondly, why's the heading written in archaizing style?

  3. Ellen K. said,

    June 14, 2019 @ 3:41 pm

    One thing that strikes me here is that even written symbols that represent a concept — namely our numerals — get used sometimes (in informal writing) to represent sounds.

    So, if a logogram represents a word, and a phonogram represents a sound, what do you call something like a numeral that represents a concept?

  4. Chris C. said,

    June 14, 2019 @ 11:37 pm

    Secondly, why's the heading written in archaizing style?

    Wherefore should he not thusly write it?

  5. ~flow said,

    June 15, 2019 @ 7:46 am

    I alluded to exactly this—"Meanings Are Made, Not Retrieved"—in a comment just days ago.

    I'm not sure where or whether I read whatever lead me to draw my conclusions, or what the current consensus, if any, in communities like linguistics and computer science about the emergence of meaning is, but at least for some time in nineties and aughts it sure did look to more than a few people concerned with the Semantic Web idea as though you just needed a sufficiently propped up XML-based markup to cast a given text into, and out would pop 'meaning' in its pure form, much like you can squeeze a lemon to obtain the juice.

    I don't think this is possible; you cannot have a writing system, a notation that in absence of a readership can perpetuate its own interpretation. Where we have reconstructed ancient scripts that were lost for centuries this has always been made possible by external archaeological and (cross-) linguistic evidence, sometimes aided by the pictorial nature of some notations (Maya, Egyptian, Oracle Bone), and certainly greatly aided by more general considerations such as commonalities of human languages and cultures. Where any kind of Rosetta stone is missing—the Phaistos Dics, Linear A, the Indus Valley script—the recorded content is likewise lost. Crucially, Ventris was able to decipher Linear B, as it records a form of Greek, but not Linear A, which records an unknown language.

    So when we write, we do so in the hope that at a later time someone—my tomorrow self, the receiver of my letter, a reader of the future—will be able to *reconstruct* enough of my marks so as to make their reconstruction sufficiently close to my original intents; we then say a successful transmission has taken place: There was a meaning m1 which I reduced to writing in one point in time, and there later came about a meaning m2, the hope being that their differential will be rather small, so that it can be said that an 'understanding of the text' has been achieved (to some degree).

    In naive materialist thinking, since there is this substance 'meaning', call it m, which is 'poured into' the text, 'cast into letters' as it were, and later does 'come out of' the text when it is (in a sometimes painstaking, in schools also sometimes literally painful process) read, deciphered—then surely, that substance m must be transported like potatoes on a cart from the field to the market.

    There is not anything in information theory that would tell us that semantics can be transported alongside with the physical bits that make up the data. Data itself is not physical but it always needs a physical substrate where it is reified as it were. Given enough technical knowledge we can from the bits extract the encoded information (and re-encode it; for example, I can make a screenshot of this page with my phone and send the picture to someone else to read; the other end may choose to run an OCR over the image to reconstruct something resembling the data that this page was built from on my computer); however, to obtain the meaning of something we need much more, including, crucially, circumstantial knowledge.

    Coming to think of it, Chinese characters with their typical two-pronged structure of a meaning-oriented and a sound-oriented part do remind one of the Semantic Web and its XML-encoded structures (where basically you annotate a given string like "2015" with an annotation saying "year of the Gregorian calendar" to nail it down). It is also quite similar to the live annotations we tend to deliver in English or any other language when we say disambiguating things a la "I love pi, I mean the number, (not the food)".

  6. Keith said,

    June 20, 2019 @ 2:53 am

    @Antonio L. Banderas

    Secondly, why's the heading written in archaizing style?

    Do you mean the heading "A B S T R A C T", with a space after each letter?

    I took a look at the PDF. There, the heading "ABSTRACT" appears with slightly increased spacing letter-spacing. The introduction of a space after each letter in the quoted text might have been deliberate on VM's part in an attempt to recreate that extra spacing, or may be an artefact of copying from the PDF and pasting into the software used to create this blog.

RSS feed for comments on this post