Thursday, February 16, 2012

Linguistics, semantics, pragmatics: words, meaning, and wacky translations

We had guests last weekend, both of them richly educated in languages and linguistics. This is what happens when you work for a university, you make friends with people who know stuff. Knowing stuff leads to conversations interesting enough to think about later, and perhaps even to blog about. Quod erat demonstrandum, or Q.E.D. as we who don't speak Latin like to say, acronymically.

A--, one of our friends, fell to describing linguistic semantics after dinner on Friday. Linguistic semantics is, more or less, the study of what people mean when they say stuff. Or write it.

This is a field of academic study, you ask? Yes indeed it is. See, deriving meaning from stuff people say is not so simple as you might think. A--, who is not a linguistic semanticist himself but does credible post-prandial impersonations, demonstrated the complexity of mapping formal meaning to spoken sentences with an example. The example he gave is situated in a classroom, a setting that comes to mind easily among those who work at universities. The statement:
Every student in this class speaks two languages.

Straightforward, right? A model of simple clarity: subject, verb, object. See? You too can play the linguistic semantics game.

But wait. What does this statement mean? Let's express it as a quasi-formal logical statement:
There are two languages, such that each student in this class speaks both of them.

Right. As in, for example, every student in the class speaks French, and every student in the class also speaks Japanese.

Or ... wait a minute ... do we mean something else by these words? Something like:
Each student in this class speaks two languages, but it is not necessarily true that each student speaks the same two languages.

That is, Peter, Bob, Jane, and Sue speak French and Japanese. Sally and Tim speak Spanish and German. Rory speaks Gaelic and Mandarin Chinese. (Me? I speak English. C'est tout. So I guess I'm not in the class.)

Which is it, then? What's meant when a person says, Every student in this class speaks two languages? Or is it the case that these words alone are insufficient to determine what's meant?

A linguistic semanticist codifies structural patterns in language that define meaning sharply from purely linguistic cues, or that lead to ambiguities like the one just illustrated.

As A-- described this business of mapping formal meaning to language -- of determining which formally logical statements are conveyed by this or that type of statement in human languages -- I grew antsy. I'm not convinced it makes sense to map formally logical statements to statements humans make in human languages. Why not? Because I'm of the opinion that very few people think in formally logical terms when they speak. In short, people aren't that precise.

I mean, have you been watching the G.O.P. presidential debates?

Theories that purport to formalize matters as nuanced as human effort to express and understand tend to strike me as reductionist. They take for granted a set of assumptions that simplify away key elements of a problem, elements that are central to the question at hand.

Take, for an example of simplifying assumptions, frictionless planes described in introductory physics courses and textbooks. Sure, frictionless planes make Newtonian laws of motion easier to construct as equations and solve, but they don't actually describe the world as it exists.

I am neither a semanticist nor a linguist, let alone a linguistic semanticist, so best to turn to the hive mind to find some more authoritative pronouncement than my own babbling on the topic of complexity and nuance in language. The hive mind, as all good intertubers know, is available to all of us on Google's home page. In response to my query, the hive mind suggests the university textbook Linguistic Semantics: An Introduction, by Sir John Lyons (Cambridge University Press, 1995). Here's what Sir John has to say:
Most language-utterences, whether spoken or written, depend for their interpretation -- to a greater or less [sic] degree -- upon the context in which they are used. And included within the context of an utterance, it must not be forgotten, are the ontological beliefs of the participants: many of these will be culturally determined and, though normally taken for granted, can be challenged or rejected. The vast majority of natural-language utterances, actual and potential, have a far wider range of meanings, or interpretations, than first occur to us when they are put to us out of context. This is a point which is not always given due emphasis by semanticists.

I couldn't have said it better myself, though I might have aimed at greater concision and a less ornate style. Put it this way: it's complicated.

For some years I worked in an administrative office at UC Berkeley called Staff Equity and Diversity Services. We were all about facilitating communication among staff and faculty across breathtaking ranges of cultures, native languages, lifestyles, class positions, educational backgrounds, and other perspective-inflecting qualities. One of our favorite buzzwords -- buzzphrases, I guess -- was shared meaning. Communication between colleagues, we believed, has to establish shared meaning to be effective and conducive to sustained, mutually respectful work relationships.

What did we mean by shared meaning? We meant the result of an involved, iterative, carefully self-conscious process by which a listener tries to understand not just what a speaker's words would mean if they were spoken from the listener's frame of reference, but what they mean from the speaker's frame of reference. This is harder than you might think. It requires that both parties do some non-trivial, time-consuming work to understand frames of reference that may be quite foreign to them.

It's a lot harder than jumping to conclusions about implied meaning.

There's apparently a sort of linguistic semantics that accounts for this approach to understanding communication via spoken and written language. It's called pragmatics. Courtesy of our other guest of last weekend, Q--, I have skimmed (I won't claim to have read, let alone fully grokked) Ruth Kempson's chapter titled Pragmatics: Language and Communication in The Blackwell Handbook of Linguistics. From that chapter:
According to Grice, who was the pioneer of the inferential approach to conversation (Grice 1975), there is a general assumption underpinning all utterance interpretation that the interpretation of utterances is a collaborative enterprise guided by a "co-operative principle" in which a speaker and hearer are engaged in some shared goal.

To me that sounds a lot closer to real-world efforts to match words with meaning.

But ... can we go back to implied meaning for just a moment? Just for fun?

Last month a Facebook friend shared a link from Gawker (thanks, Elliot!). It summarized a post by Jake Adelstein on the Japan Subculture Research Center's website. The post featured photos snapped by Zarina Yamaguchi at a department store in Osaka; the post was titled It’s no ordinary sale. It’s a FUCKIN’ SALE! The image says it all (at right).

Those signs in the department store beg a certain question, don't they? What (on Earth) did the marketing department mean when they characterized storewide discounts of 20% as a Fuckin' Sale?

Inquiring students of linguistic semantics want to know.

Related posts on One Finger Typing:
Google Translate, AI, and Searle's Chinese Room
Music, memory, nostalgia ... and the novel
Craft and art: erasure and accent
Aleksandar Hemon on Narrative, Biography, Language

No comments:

Post a Comment