L. J. Kramer, "You gotta love a guy who's studying the root of 'tree', for elsewhere would you start?" Or words.
Here's some more on slightly related stuff: the diachrony of implicature.
Recall that for Saussure, 'arbor' is the acoustic image of the concept, 'tree'. Any implicature that was in the _past_ is no longer synchronically relevant, Etc.
See how a Gricean may refute all that, and fail
--
E. Traugott & R. Dasher, _Regularity in Semantic Change_. Cambridge University Press, ISBN 0521583780, xx+341p. Cambridge Studies in Linguistics 96.
(Reviewed by D. Lewis, University of Oxford).
"_Regularity in Semantic Change_ (RSC) focuses on the semantic-pragmatic interface. Traugott and Dasher (T&D) defend the position that there are cross-linguistic unidirectional tendencies in semantic change, at least in certain domains, and that internal semantic change largely occurs as the conventionalization of implicature."
"A further claim is that speakers/writers are the key innovators (implicatures are
controlled by the speaker/writer), hence the name of the theory proposed:
the 'Invited Inferencing Theory of Semantic Change' (IITSC)."
"The major type of semantic change is claimed to be subjectification."
"The emphasis is on
"meaning changes that are primarily linguistic
and that have implications for constraints on
lexical insertion or grammatical function" (p. 11)."
"Evidence is presented from the areas (broadly construed) of modality and deixis."
"The historical developments of modal verbs, discourse markers, performative verbs and social deixis are charted."
"Claims for unidirectional tendencies in semantic change have frequently been made in studies of grammaticalization. Traugott, in previous work, has examined the development of English discourse markers from a grammaticalization perspective (e.g. Traugott 1995)."
"RSC argues that although
"the greatest degree of semantic regularity has
so far been found in conceptual structures the lexemes
of which are typically associated with grammaticalization"
(p. 3),
regularity in semantic change is not limited to grammaticalizing lexemes."
"Regular patterns of semantic change are also found in other domains,
"especially lexemes that are verbal and
(in relevant languages) adjectival or adverbial"
(ibid.)."
"T&D do not dwell on the likely distinctions between the typical developments in these categories and those in nouns (or why nouns might be "particularly
susceptible to extralinguistic factors" (p.4)), but they keep the focus of this book firmly on verbal and adverbial development."
"The IITSC is set out in the first chapter. Assuming a broadly cognitive view of language, and drawing in part on ideas from prototype theory and construction grammar, the IITSC aims to account for the semanticization of pragmatic implicatures."
"The ITTSC focuses on the fact that "online-production and processing make
use of essentially syntagmatic relations and associations" (p. 9), and posits that
associative, metonymic relationships are more important in change than
metaphorical, analogical ones."
"In simplified terms, the model has speakers/writers exploiting invited inferences and re-weighting implicatures to the point where a lexeme acquires a stable _utterance-type meaning_ (Levinson 2000), i.e. a default, though defeasible, interpretation in a context type."
"This utterance-type meaning can then semanticize into a new sense, Meaning(2), alongside Meaning(1)."
"The path is:
From
i. coded meaning
to
ii. utterance-token meaning
to
iii. utterance-type (pragmatically polysemous) meaning
to
iv. new coded meaning (and so semantic polysemy).
"The mechanisms of such changes are said to be language-external, in that they are processes of reasoning by speakers/writers (p. 40)."
"There follows (ch. 2) an overview of prior and current approaches to semantic change, from Breal's still-influential categories of pejoration, amelioration, contagion, etc. through to recent work in historical pragmatics."
"From the early twentieth century, T&D draw particular attention to work on change within semantic fields."
"From more recent research, they focus on analyses of metaphor and metonymy by scholars of grammaticalization, on studies of subjectification, and on the formulation of neo-Gricean pragmatic principles and the relevance of these to semantic change."
"In particular, they build on Horn's neo-Gricean principles to argue for a 'Quantity-heuristic ('make your contribution sufficient', implying 'at most p'), a Relevance-heuristic ('say no more than you must', implying 'at least p'), and a Manner-heuristic ('avoid prolixity/marked expression-marked situation')."
"It is application of the Relevance heuristic, they suggest, that can result in semantic change of the type discussed in this book."
"The development in some modal verbs of epistemic meanings from deontic meanings is described in detail in the next chapter. T&D present case studies of the development of English 'must' (from ability/permission through obligation to epistemic uses), English 'ought to' (from possession 'have' through obligation to epistemic uses). These developments evidence a tendency towards more speaker-oriented meaning, and so greater subjectification. The forms' acquisition of modal meaning also involves
acquiring 'procedural meaning' in addition to 'content meaning'."
"The next case study is the development of adverbials with discourse marking functions. Analyses of 'indeed', 'in fact', and 'actually' show how the semantic development of each (manner or similar meaning -> epistemic meaning -> elaborative or clarificatory connective) is paralleled by its syntactic development involving ever-increasing scope (VP-internal adverbial -> sentential adverbial -> clause-external discourse marker)."
"'Well' and 'let's' then further exemplify the development of or extension of intersubjective meaning."
"In all these cases, subjective and often intersubjective meanings develop out of more objective, 'content' meanings."
"Chapter 5 describes the development of performative verbs and constructions from non-performatives. Typical sources for performative verbs are terms relating to visual perception, vocalization, mental states and object manipulation. Detailed histories are given, again emphasizing subjectification of meaning, of English 'promise'.
"The last main chapter deals with the development of social deictics."
"Regularities in semantic change are hard to pin down. Ullmann described how, following Breal, in the 1880s-1930s period, scholars set out to discover laws of semantic change, and to establish taxonomies of change. But "the quest for 'laws' met with very limited success, and the classificatory zeal resulted in a number of ambitious schemes built on slender empirical data" (Ullmann 1962: 196). Slender empirical data can still be a problem."
"The past two decades, however, have seen a renewal of interest in both semantics and language change. Typological studies and grammaticalization studies have both provided an impetus for a new look at the possibility of universal pathways or tendencies in meaning change. RSC can be seen as a product of such impetus."
"In past discussions of semantic change, perhaps too little attention has been paid to
(a) the notion that different types (entities, attributes, predicates) or semantic domains may tend to undergo different kinds of change by different kinds of mechanism, and
(b) the relevance to semantic change of the context types (both textual and
communicative) in which lexemes regularly occur."
"RSC takes both into account."
"Moreover, many of the difficulties inherent in interpreting the sorts of data with which historical pragmaticists have to work are acknowledged and discussed."
"RSC assembles an extremely valuable range of case histories of lexical semantic change and builds a persuasive argument for the importance of the role of discourse context in semantic change, and for gradual metonymic extension."
"One objection to the IITSC might be that the theory is not properly predictive. But such predictability is not the aim. As Harris & Campbell point out, "That the fact of change is not fully predictable does not entail either that change is random or that the limits of change cannot be stated" (1995:6). The claims of RSC are about what kinds of internal change are most likely to occur, should change occur, and by what mechanisms."
"Internal (cognitive, psycholinguistic) and external (socio-political) pressures for change may conflict. The main claim is that semantic change is not random, but is subject to identifiable regular pressures which, when they prevail over other, ad hoc pressures, lead to greater subjectivity of meaning, by the gradual semanticization of pragmatic inferences resulting from speaker/writer intention."
"Another, more substantive, possible objection concerns frequency. Claims about regularities in change are necessarily statistical claims. The unidirectional argument is an argument about the relative frequency of particular semantic pathways, yet the statistical significance of the changes discussed is not addressed in RSC. Ultimately, for such generalized claims about change to be upheld, it will be necessary to clarify what the semantic change population is and what sampling is
appropriate. This difficulty is perhaps a weakness of the semasiological approach. The data are persuasive, but by electing to examine the histories of small groups of 'successful' expressions belonging synchronically to certain modal, subjective areas of meaning, and without any quantification, it is hard to reach firm conclusions about semantic change in general. The same applies to the claim that, at the level of individual lexemes, it is 'preferred strategies' of speakers/writers that lead to semantic change, since this is presumably a claim about frequency of strategy."
"Occasionally, the reader feels that categories are in danger of becoming blurred. Four main pragmatic-semantic diachronic regularities are proposed in RSC."
(1) -subjective -> +subjective
(2) contentful -> procedural
(3) increase in scope
(4) +truth-conditional -> -truth-conditional.
"This implies a semantic theory that posits at least these four parameters of meaning. However, T&D do not claim they are necessarily independent of each other (p. 284), and in fact their status and inter-relations warrant further investigation and clarification."
"'Procedural meanings', for example, are described as "primarily indexical of speaker/writer's attitude to the discourse and the participants in it; they index metatextual relations between propositions or between propositions and the non-linguistic context" (p. 10)."
"But it is not quite clear what the evidence is for the binary distinction, nor whether contentful and procedural meanings are assumed to have different cognitive qualities, nor exactly how 'procedural' relates to 'metatextual' or 'subjective' or even to 'pragmatic', with which it seems sometimes to overlap (e.g., 'contentful meaning' is also contrasted with 'pragmatic meaning' (p. 96))."
"Overall, it is not immediately obvious that the contentful/procedural distinction is necessary or useful to the main arguments of RSC."
"There is also some uncertainty over the status of subjectification, which is described in the conclusion as "the main mechanism of semantic change" (p. 279), when previous chapters had seemed to argue that subjectification was a type of semantic change."
"We need many more detailed, quantitative analyses across time of lexical tokens in their textual and communicative contexts. This is a task that future studies in historical pragmatics, using large historical corpora, should have much to contribute to. Meanwhile, RSC is a most valuable contribution towards addressing this vast gap in our understanding of language, and towards a better understanding of the complexities of lexical semantic change."
References:
Harris, A & L. Campbell. Historical Syntax in Cross-Linguistic Perspective. Cambridge: Cambridge University Press.
Levinson, S. Presumptive Meanings: The theory of generalized conversational implicature. Cambridge, MA: MIT Press.
Traugott, E. The role of the development of discourse markers in a theory of grammaticalization'. Paper presented at the Twelfth International Conference on Historical Linguistics, Manchester, August 1995.
Ullmann, S. Semantics. Oxford: Basil Blackwell.
Subscribe to:
Post Comments (Atom)
Posts
When I was first exposed to Peano's postulates, I was struck by how logically similar they seemed to Euclid's axioms. That may have been the day I became a reductionist. Anyway, a reductionist I am, and I pretty much see things through a very small number of windows, most notably these:
ReplyDeleteNatural selection
Diminishing returns
Second Law of Thermodymanics
For example, I speculate that language evolves to be only as inflected as necessary to strike the most survivable balance between the effort to learn and the time saved vs. periphrastics. The tendency would be for inflection to be added as it becomes too much work to say all of the words in a periphrastic and are then tested by entropic erosion, which sticks or not.
English could have separate words for "he eats" and "she eats," but we don't. We do, however have inflected pronouns for "it the female" and "it the male." But we don't need to have them. "It the male eats" is perfectly serviceable, but it's too much work. Yet, we say "he eats" and "she eats" when "eatso" and "eatsa" could be recruited to mean those things.
Natural selection may explain where these lines are drawn. It's interesting that change arises from the unwillingness to expend energy (and energy is the only antidote to entropy, which is where evolution and the Second Law intersect). Whether that energy would otherwise be spent on the capital cost of learning an inflection or the current cost of using a periphrastic can determine the direction of the linguistic change toward ro way from inflectedness, but the change in either direction can be seen as the result of a consistent strategy of expending as litle energy as necessary.
The processes described in the post above, being essentially psychological, must reflect an evolved strategy of human beings for survival, i.e., a specific way of developing languages and modifying their semantics. Thus, a high level of cross-linguistic patterns should be discernible.
With Girce's maxims in mind, consider these rules for getting from 0 to 10:
1. Use positive numbers. Negative numbers are unhelpful.
2. Add as much as you need, but no more than 10.
3. Use numbers. You cannot get from 0 to 10 by adding letters.
4. Record each number so that it can be verified.
That's why, from an impertinent reductionist layman, Grice's maxims elicit a heartfelt "well, duh." Of course, that's how language works - it's how everything works. Linguistics thus seems to me a branch of psychology, which is a branch of biology, the root, so to speak, of the tree.
Experience has taught me that someone has said what I'm saying. I've learned from my exchanges with JL, and from many years of speculation, that I never say something that JL cannot show me someone else saying. So please consider this comment just another request for references and refutations.
Never say never, Larry. Of course I love you and none of them name-dropping droppers of name I sometimes add to justify my existence. Etc.
ReplyDeleteKramer reads Grice's maxims and writes:
ReplyDelete"duh"
implicating,
"Of course, that's how language works - it's how _everything_ works."
Vide, The theory of everything. Or is it "_A_ theory of everything"?
I'm not sure it's how everything works though. Grice, on his way to Oxford, wrote on a plane (Chapman notes), "I should read _Selfish Gene_". We don't know if he did. It seems to me that some things work for Grice
natura
Otherw work
contra natura
--- and others don't. ("Natura morta", as I heard that beautiful phrase in Sophia Loren's "Nine" -- now playing).
So there may be a scheme of how things _work_. Then there's expectations that things _will_ work like that. But then there's realizations of these expectations in, say, people. And people have their pre-rational parts ('contra natura' sometimes) and their rational parts.
So even if _everything_ works as Grice says (via the maxims), we still need to perhaps spend some time in consideration how we _flout_ the things that work. And why the flouting of the things that work is still _natura_ rather than _contra natura_. Etc.
I think Grice saw the history of the English Language as
-- pretty silly in parts.
-- unphilosophical
-- totally empirically, and guided by
irrational forces.
Thus, his examples are "anecdotal". People like Traugott, on the other hand, make their living out of this.
One good difference here is that for Grice, and Davidson, Languages, strictly, does not exist. As P. Suppes writes in PGRICE: it's an abstraction, and one of the worst of them at that, in being dangerous. There's speakers, rather than language. Davidson, in "A nice derangement of epitaphs", defines a language as a median for speakers, or something.
The contigent, unphilosophical character of much of the so-called 'language evolution' (which I call 'decay' just to tease Sperber and Wharton) is illustrated by this jocular side comment by Grice. In studying why words should stick to one sense at most:
He is considering the idiotic 'development' of the word
animus --> anima --> animal
English, 'animal'
"'Animal' can mean, in contemporary English, either i or ii."
i. member of animal [sic] kingdom.
ii. beast.
"'Animal' infringes a weak principle to the effect that a further 'sense' should NOT be recognised if, on the assumption that 'animal' were to have a specificatory further sense, the identity of the sense would be PREDICTABLE. For it could no doubt be
predictable that IF 'animal' were to have such a sense, it would be one in wwhich the word did NOT apply to human beings."
"But it seem NOT to be predictable
-- HISTORY OF LANGUAGE APART --
that anyone would in fact use 'animal' to mean 'beast'.
(He is postulating the _arbitrariness_ in the semantic history of 'animal' to the genealogy of disjunction, where, "given a truth-functional 'or', it IS predictable (assuming conversational principles) that people WOULD use "p or q" to imply the existence of non-truth-functional grounds [for the uttering of "p or q". Etc. (WoW:iii)
Excellent comment, JL.
ReplyDeleteI have been trying to tease out statements like Suppes's by saying things like "language is a behavior" or, maybe more precisely, the reification of a behavior. That's why language seems to me so natural a window into evolutionary psychology. (I was confident such statements were there and that you would have access to them.)
I doubt hat Grice would have needed to read The Selfish Gene. You and I have both read some silly rants opposing Dawkins's contentions, but they are, indeed, silly rants. Dawkins's pronouncements are self-evident and tautological: genes that reproduce persist, so any trait X that enables carriers of the coding for trait Y to reproduce will cause trait Y to persist.
Dawkins's contribution is to recognize that traits X and Y are logical devices for which the same physical device may or may not serve. Thus, clotting blood is both trait X and trait Y, but, in a genetically homogeneous population, altruism can be trait X and all of the other traits of all of the altruist's relatives trait Y. That's the book. Everything else (and some of what I wrote above) is commentary. I suspect Grice might have said "Duh! Of course evolution works that way. Everything works that way." (My kind of guy, that Grice.)
I'm not sure what you mean by "flouting" things that work. Bad faith interactions flout things that work. But otherwise, it seems to me that apparent floutings simply (or complexly) invoke something else that works:
He drew a circle that shut me out -
Heretic, rebel, a thing to flout. But love and I had the wit to win;
We drew a circle that took him in.
Edward Markham.
Excellent comment, too. I was fighting with the italicisation of your "duh", but had failed. Etc. Apparently, I used to receive the OED bulletin, and one recent copy had the note: new entry in the OED: "duh". Apparently, the first quote they were able to trace was Homer Simpson. Anyway, T. Wharton -- our tim w -- who knows you well after you rediscovered the pseudocodes for him -- is studying interjections. I offered myself to list them alphabetically. Indeed, I don´t think we had others under the "d", so your "duh" welcomed. ("I never say a big big D", of Pinafore, does noit really count in that it means, "condemned shall I be if I believe this stuff").
ReplyDeleteYes, the comment by Grice was brilliant: "history of the language apart". It´s like saying: the universe is Newtonian, everything Einstein proved wrong about him apart."
You are right that Grice would have been seen to utter "duh" on occasion. And I´m pleased you liked the Suppes connection. This man was a genius, Suppes. He taught at Stanford for _years_ and I was surprised to see his name in the PGRICE festschrift. Davidson, on the other hand -- I´m currently studying this phrase, "on the other hand", elsewhere -- is your regular´one.
Recall that Grice is a Nominalist with a capital N. For him it all begins with a one-off token of an utterance. An x. Then this gets classified as a type, X. It grows a capital X. This is all we need to say an Utterer U means by uttering x, that p. THEN we may care to ask if ¨x¨ or ¨X¨ means something. Within the utterance, Grice considers "utterance-parts". I´m not sure what he meant by this, but my Grice´s Shaggy Dog story illustrates it. He is meaning that in some contexts,
She said that he was a clown
"...was a clown"
does not really count as an "utterance". Only an "utterance-PART". Following Strawson (his student actually) he focused on utterance parts that REFER and utterance parts that PREDICATE.
"Words" are utterance-parts, and sentences are TYPES of utterances -- but only in a very abstract, higher level of his programme would he speak of S -- sentence -- belonging in language L.
I always adored that nominalism. Even if wrong, it´s as they say my cup of tea.
Etc.
I think Homer's interjection is transcribed as "Doh!" It's quite different from "Duh." "Doh" means "I'm such an idiot"; "Duh" means "You're such an idiot."
ReplyDeleteI would have thought that everything that isn't the whole utterance is potentially an utterance part. Recall my "thingies" that consist of more than one word but mean something that does is not the same as the words themselves. An Anglo/Norman legalism to such as "to have and to hold" offers an example: it is the phrase, not the words, that has the legal consequence of transferring title. (Chew on the choice of verb - "has" or "have" - in that sentence.) So the phrase is an utterance part. N'est-ce so?
Doh, duh. Or duh, doh. Etc. Yes, you are right. I suppose the OED3 already lists 'doh'. Anycase, we have two under 'd' then, pretty reversible in parts, too. I may be providing a blog entry after this on the old flouter quoting your excellent Markham. Yes, the title of Grice's vith William James goes by the clumsy title: utterer's meaning, sentence meaning and word meaning. Foundations of language. I don't know if Grice took the
ReplyDeletetotal
vs.
part
distinction seriously. It seems, indeed, there are some problems here.
I like to take an 'utterance' as any vehicle for meaning-nn. In this sense, utterance gets defined in terms of meaning. A stone is not an utterance, but my throwing it is. As R. B. Jones recently commented THIS BLOG, Wittgenstein's Fork. It was not his fork, it was his flourishing of the fork -- that was all his own.
The fork may MEAN (e.g. if you leave it somewhere for someone to realise that you'd been there), etc.
Anyway, back to utterance parts. Grice has HW, for example, hand-wave, as a favourite example of his.
I don't think he went into the structure of utterances per se.
It would be a categorial mistake to say that an utterance has a subject and a predicate, for example. An utterance refers to a 'phonic' act in Austin's parlance. A referential and predicative structure belongs to the 'rhetic' act, as he'd also say, Austin.
In the case of the shaggy-dog story,
Plato, my dog, is shaggy.
So we say that '... is shaggy' is part of the whole utterance.
Utterance is perhaps the wrong expression. It does not yield to a good latinism. Utteratum sounds clumsy. Etc.
In any case, it may do to play grammatically or etymologically, or Grecianly, with pars and totum.
Meriology is the study of parts. Right. Holism would be the study of the totum. Perhaps something along these lines can be built for Grice's reflections there. Will think about them.
When I wrote my PhD I concentrated on a special type of utterance-PART which is neither referential nor predicative. But it is not one of a BASIC procedure that I wanted to analyse, but RESULTANT procedures.
"and", "or", "if" and "not".
So take "and". We have the utterance-whole:
Peter is married and Mary is married.
(I'm studying reciprocals with Bayne)
But we can abstract two utterance parts here
S1
S1-a & S1-b
i.e. a conjoint or co-ordinated structure. Now, 'and' merits a specification in terms of meaning with some level of independence of what "Jill is married" or "Jack is married" may mean. This is the truth-table for 'and'.
The truth-conditional or truth-functional meaning of 'and' is all that the resultant procedure for the utterance-part indicates.
The problem then ensues as to the meaning
"Jill is married and Jack is married but not to each other" or
"Jill married and Jill had a child"
Surely no "and then" is implicated _in abstracto_. But Grice challenged Strawson. For Strawson had said that "and then" is PART of the meaning of the utterance-part, "and". And Grice wanted to keep the truth-functional meaning _bare_. So, by applying
UNIGUITY plus Implicature
(or Grice Saves, but there's no such thing as a free lunch), we postulate a 'pragmatic rule' as Strawson 1952 called it,
'be orderly'
and abra cada bra. We can keep the truth-functional meaning of the utterance-part "and" as essential, and add another resultant procedure which gets applied to this truth-functional resultant procedure to the effect that: On occasion, when an utterer utters "S1 and S2" in a context where the maxim, be orderly, is felt as operative, the utterer may be claimed to have meant, "S1 and then S2".
Etc. But now to flouting.
abra cada bra.
ReplyDeleteI like that. Open every bra. Sounds more like an undergraduate project than a Ph.D. thing though.
Yes, hocus pocus.
ReplyDeleteOf course, my PhD thing _was_ an undergraduate thing, but don't spread the word!
You should SEE the types of seminars that were required from me to be able to find myself in a position to be allowed to be experienced in the ability to be able to produce the 'sotto voce' defense of my PhD thing!
I never had a REAL interest in the 'and' thing -- only Grice. But one required seminar was all about
Susan Haak.
Gregorio Klimovsky was giving it. I had to PASS the seminar, write a term-paper, defend my term paper in front of other doctorands (sic), etc.
And I already had proposed a plan for my dissertation which was prety easy:
"My thesis will focus on Grice. I want to introduce Grice. And then I'm going to develop some theory of Grice. Finally, to conclude with Grice."
---
So, I had to swallow this BORING book by Haak, and tr. into the vernacular, which didn't help. And see how I could connect it with Grice.
So GRATEFULLY, or gricefully, Haak cares to mention Grice -- and so I used that as the key.
My term paper for that PhD seminar I entitled, "Post-Modernist Grice". The joke fell flat on Gregorio Klimovsky -- R. I. P. -- he died last year --, and he gave me an A+.
For that paper, which was later incorporated as ch. ii of my PhD -- tricks PhD students play --, I made some wise decisions:
-- avoid "not". After all, it's not a connective. So why bother? Stick with the truth-functional connectives.
-- avoid 'iff'. It's a reiterative connective.
-- avoid 'if' if you can. Hey, Grice is not sure his account can save the truth-functional side to the Philonian conditional.
-- This left me boringly with things like
"She got pregnant and got married"
and
"She was in the garden or in the kitchen (last time I saw her)."
THAT I had to write on, with a level of fluency and moral integrity to get the A+, which was the only thing I cared for.
So I came up with some strange symbolism, like
pdf, --> frv gd, of the part
p & q -- and then. pdf, utterance-mean.
etc.
Very impressive.
In an earlier stage of my PhD (which I must have, somewhere), I was ambitious enough to get from 'iff' to 'all' and 'some' and cover all of Horn's scale-implicatures.
Some hope!
I soon realised that 'some' and 'all' are NOT truth-functional. So why bother? Introducing the topic of non-truth-functionality would have meant complicating my brain and my tutor's brain.
Plus, while opening every bra was never much of my thing, I had _THINGS_ to do, and surely while Grice saves, he's not _ceteris paribus_ going to save _your_ life. Etc.