conversation and syntax

Tom Givon tgivon at
Mon Jun 9 21:33:19 UTC 2008

I think we are finally getting a discussion of interesting issues, so 
again we should thank Fritz for initiating it, by whatever means. This 
once again brings home the fact that ideological homogeneity is not 
necessarily good for intellectual growth. Now,  Fritz raises two issues 
that maybe I could respond to.

GRAMMATICALIZATION: What does it really mean (quote) "...[All that I 
have ever said in print is that] what is called 'grammaticalization' is 
the by-product of a number of interacting processes..."? Does this mean 
that there are many different communicative functions that get 
grammaticalized, and that the (by-)products  of these multiple processes 
of grammaticalization are different morpho-syntactic constructions? If 
that's all it means, there is nothing earthshaking about it. I'll 
subscribe any time. On the other hand, it could also mean "there are no 
(strong!) general trends that characterize the diachronic rise of 
morpho-syntactic construction" (as the Campbell/Janda/Jacobs crowd would 
have it). This is is a substantive negative claim, one I don't see how 
one could accept. Having studied grammaticalization for ca. 40 years 
now, I'll have to take a strong exception to such an interpretation. 
What those of us who have spent a lifetime trying to understand the 
genesis of grammar (diachronically, ontogenetically, phylogenetiuvcally) 
would say is that there are VERY STRONG general tendencies--and 
principles--that characterize such a developmental process. Whether one 
would want to call them "theory", "rules" or "laws" is a matter of 
utter  indifference to me. I would be satisfied with "high degree of 
generality" or "high degree of predictability". As the late Ernst Mayr 
said long ago, biological regularities are not like the laws of physics. 
They emerge out of multi-factored, complex, adaptive environments, where 
often adaptive factors clash with each other. So even the strongest 
regularities are less that 100% ("generative"). This has never 
discouraged biologists from seeking powerful theoretical explanations. 
Why should it discourage linguists?

MAIN STEAM FUNCTIONALISM:  Here is the other quote from Fritz: "...Where 
I part company with 'mainstream functionalism' (if there is such a 
thing) is over the question of whether performance factors (including 
frequency) are stated in the grammar itself...".  First, I have no idea 
what Main Stream Functionalism is. For the record, I have never accepted 
the position that performance factors, such as frequency, are "stated in 
the grammar itself". They have, obviously, to be stated somewhere in the 
theory that explains how grammars arise, through the three developmental 
trends we know. So if we accept that in order to understand how "the 
grammar itself" arises diachronically, ontogenetically or 
phylogenetically,  we must consider "performance factors", I would be 
quite happy to sign up on that, having spent a lifetime trying to 
understand just that. Once again, I would like to suggest an analogy 
from biology. In biological evolution, the interaction of adaptive 
behavior ("performance factors") with random genetic mutation is what 
actually controls adaptive selection--thus the emergence of new 
structures.  Mutations don't do the trick by themselves. So there is a 
strong precedent for "performance factors" at least contributing to the 
genesis of structures. (If you want a more authoritative source than me, 
look up an article by Fernald & White (2000) in M. Gazanniga (ed.) "The 
New Cognitive Neouroscience", Cambridge:MIT Press).

PS:  It is a pity that FUNKNET does not allow attachments, because I 
have 4 (two diachronic, two child language) chapters strongly bearing on 
this discussion. I'll send them privately to anybody who expresses 
interest. They are also available on the Rice-lingfistics-12th-symposium 
website, I'm told).

Peace,  TG


Frederick J Newmeyer wrote:
> First, I would like to thank those of you who defended the 
> appropriateness of my posting and/or provided substantive comments on 
> the topics that I discussed in the paper.
> Let me start by calling attention to what Ron Langacker has called the 
> 'Rule-List Fallacy'. Ron noted, completely correctly in my opinion, 
> that it was a fallacy to assume that lists have to be be excised from 
> the grammar of a language if rules that subsume them can be 
> established. The converse of this fallacy is equally fallacious: that 
> rules have to be be excised from the grammar of a language if lists 
> can be established. Even if it were the case that a huge percentage of 
> language users' output could be characterized by lists (formulas, 
> fragments, etc.), that would not exclude their also have a grammar 
> composed of rules (or their notional equivalents) that allow hearers 
> to analyze unfamiliar collocations and assign to them structure and 
> meaning.
> I have a couple of comments on Tom's useful posting of earlier today.
> 1. Tom says that 'Fritz does not believe in grammaticalization'. I 
> can't imagine what it would mean 'not to believe in 
> grammaticalization'. All that I have ever said in print is that what 
> is called 'grammaticalization' is the by-product of a number of 
> interacting processes. There is no historical dimension to my 
> analysis, it is true, but that is because I endorsed (and briefly 
> outlined) Boye and Harder's account of the grammaticalization of 
> complement clauses, which brings in historical developments. By the 
> way, it is not always the case that in grammaticalization 'semantic 
> change precedes syntactic readjustment'. For examples of the opposite 
> order of events, see chapter 5 of my book 'Language Form and Language 
> Function'.
> 2. For quite a few years now I have endorsed the position that 
> grammars are to a large degree functionally motivated and that 
> frequency is an important factor shaping them. Where I part company 
> with 'mainstream functionalism' (if there is such a thing) is over the 
> question of whether performance factors (including frequency) are 
> stated in the grammar itself. I think not, though, obviously this is a 
> complex and difficult question. I did not take up this question in my 
> paper because I believe that one can make a strong case that 
> sentential complements are syntactically subordinate and that grammars 
> are far more than 'fragments' and 'formulas' without addressing the 
> question of where the dividing line between competence and performance 
> might be. That's a topic for a different paper.
> Best,
> --fritz
> Frederick J. Newmeyer
> Professor Emeritus, University of Washington
> Adjunct Professor, University of British Columbia and Simon Fraser 
> University
> [for my postal address, please contact me by e-mail]
> On Mon, 9 Jun 2008, Tom Givon wrote:
>> Dear FUNK folks,
>> I think Fritz's article has already achieved something, at least 
>> potentially, that has been absent from FUNKNET for quite a 
>> while--serious substantive discussion of a relevant topic. So he's 
>> done us a service, however resentful of him some folks may be. I go 
>> with Barbara King in noting that the full article is available, and 
>> it is better to react to it rather than to the abstract. Since I have 
>> seen that article a couple of months ago, & read it in full, I think 
>> maybe I can share with you my (very brief) take on the issues Fritz's 
>> raised. This comes out of private correspondence with Paul Hopper, 
>> who berated me for not jumping to Sandy's defense. Here goes:
>> Well , in my innocence I thought there were at least two empirical 
>> arguments in Fritz's article. (i) Semantic: It is not true that ALL 
>> uses of V-COMP constructions are the "grammaticalized" ones. Some are 
>> not. And (ii) Syntactic: The uses that are semantically 
>> "grammaticalized" are not necessarily syntactically one-clause--yet. 
>> Syntactic properties have to be demonstrated by syntactic tests, 
>> independently of semantics. This, by the way (unnoticed by Fritz) is 
>> a run-of-the-mill observation in grammaticalization: Semantic change 
>> precede syntactic re-adjustment. Tho of, course, Fritz does not 
>> believe in grammaticalization, so he couldn't see this part of his 
>> own argument.
>> What was missing from Fritz' article were the developmental 
>> ("emeregent") components, which are well backed up (I think) by 
>> frequency counts:
>> (iii) That in diachrony & ontogeny (child language acquisition), 
>> V-COMP constructions first emerge is the "grammaticalized" (direct 
>> speech-act) use, and only later develop the other (two-clause) use 
>> (Diessel 2005). This omission exposes Fritz's lack of interest in the 
>> role of frequency distribution in the "emergence" of syntax, indeed 
>> his lack of interest in development/emergence. And
>> (iv) The primacy of spoken language in these two 
>> developmental/emergent processes, as demonstrated by frequency 
>> distribution (in this case, of the "grammaticalized" sense). Both of 
>> these omissions are due to Fritz' abiding faith in "competence", and 
>> his disinterest in the role of "performance" (in this case, rising 
>> usage frequency) in creating "competence" (syntax). But Fritz's 
>> attitude is echoed by that of some confirmed functionalists, who 
>> believe only in emergence but not in any RELATIVELY stable product of 
>> emergence. So all in all, Fritz' article is indeed an interesting 
>> exemplar, and people could benefit from seeing its internal 
>> contradictions, or at the very least, it's lack of explanatory ambition.
>> The last comment I have goes to Paul's demonstration of usage 
>> frequencies from Google. What I missed in his demonstration is an 
>> explanation of what these frequencies mean. Frequency counts are only 
>> meaningful in CONTRASTS: "X is frequent, as against Y that isn't". 
>> And if not in x-y contrasts, than at least against some BASE OF 
>> COMPARISON--"what is the total population within which X appears a 
>> certain number of times?" So frequencies by themselves are not all 
>> that meaningful. Still, if you have a hypothesis you want to test, 
>> you can formulate a frequency count in a way that would be 
>> meaningful--and test your hypothesis.
>> This is, by the way, an argument I had with Andy Pawley when he came 
>> out with his formulaic usage papers from the 1970s and 1980s. The 
>> phenomenon was surely recognizable. What I wanted to know was (a) 
>> What was the frequency distribution of formulaic vs. non formulaic 
>> usage in particular texts. Or (b) in what communicative contexts do 
>> we have higher vs. lower frequencies of formulaic usage. Andy wasn't 
>> interested in frequencies at the time, so I talked one of my grad 
>> students, Lynn Yang, into doing her MA thesis on this issue. It was a 
>> two-part experimental study, one with two contrasting videos 
>> ("Chicken story" vs. "Breakfast Story", a contrast of cultural 
>> familiarity) the other with recorded coffee-shop chats under two 
>> contrasting conditions (familiars vs. strangers). The results were 
>> quite instructive. That MA thesis may be available wherever U. Oregon 
>> theses are available (somewhere in cyberspace, I guess).
>> Best, TG
>> =====================
>> Paul Hopper wrote:
>>> Dear Colleagues,
>>> In Fritz Newmeyer's article on conversation and syntax, he gives the 
>>> following example of a text in which an analysis in terms of 
>>> formulaicity and formulaic fragments would be impossible:
>>> A: hi  B: hi so did you hear what the topic is
>>> A: yes it's about terrorism right  B: yeah  B: um  A: so what are 
>>> your feelings on that [laughter
>>> B: i have [laughter] i personally can't imagine anyone staying calm 
>>> [laughter]
>>> A: yeah nor can i yeah B: um you would even i- though if you're 
>>> panicked i would assume you would try and  B: keep your head clear 
>>> enough to act to protect yourself but  A: right  A: yeah i don't 
>>> know if there was an explosion or something i don't it it's a shock 
>>> so i don't know that anybody can really think about it and
>>> control themselves B: um  B: right even with all the um  B: (( 
>>> [sigh] the ))  B: the publicity and media coverage you know that's 
>>> been on that topic A: (( [mn] right ))
>>> B: twenty months it's still um  B: is something that you wouldn't 
>>> be  Newmeyer writes (MS p. 13): "There are certainly formulaic 
>>> expressions here: hi, right, take in stride, I don't think, and 
>>> possibly a few others. But in other respects the transcript reveals 
>>> a sophisticated knowledge of syntax that defies any meaningful 
>>> analysis in terms of 'fragments'. The speakers know how to handle 
>>> purpose clauses, wh-inversion, relative clause attachment, 
>>> participial complements, and much more. If these are somehow to be 
>>> subsumed under the rubric of 'fragments', then I would say that this 
>>> infinitisemally small sample of natural speech would have to contain 
>>> at least two dozen fragments. How many more would be needed to 
>>> describe a typical speaker's daily output?"
>>> Regarding Fritz's last statement: Dwight Bolinger said somewhere 
>>> that there's a reason the human brain has trillions of cells...! 
>>> Well, I identified the groups of words that I would suspect are 
>>> formulaic and typed them into to see if they were as 
>>> unique (and therefore syntactic) as Fritz claims. The results are, 
>>> it seems to me, consistent with the idea that the speaker is indeed 
>>> stringing together formulaic fragments. The statistics are Google's, 
>>> of course--they are rough and may fluctuate with different trials.
>>>   - Paul
>>> -----------------------------
>>> did you hear what the topic is\ "Topic" doesn't occur, but about a 
>>> dozen other NPs do; the formula is: \did you hear what the * is\.
>>> it's about terrorism\ occurs 5,240 times
>>> what are your feelings on that\ occurs 266 times
>>> can't imagine anyone\ occurs 387,000 times; staying calm\ occurs 
>>> 275,000 times
>>> nor can I\ occurs 1,210,000 times
>>> you're panicked\ occurs 2,090 times
>>> i would assume you would try and\ occurs twice
>>> to keep your head clear enough to\ occurs 14 times.
>>> to act to protect yourself\ occurs 14 times
>>> if there was an explosion or something\ occurs 6 times (!)
>>> it's a shock\ occurs 98,400 times
>>> i don't know that anybody\ occurs 17,900 times can really think 
>>> about it\ occurs 69 times
>>> control themselves\ occurs 406,000 times
>>> even with all the publicity\ occurs 190 times
>>> media coverage\ occurs 295,000 times
>>> on that topic\ occurs 2,970,000 times
>>> in the last twenty months\ occurs 1,070 times [NB specifically with 
>>> "twenty", not just any number!]
>>> something that you wouldn't be\ occurs 140 times
>>> able to take in stride\ occurs 289 times
>>> -----------------------------------------------------------
>>>> Dear Colleagues,
>>>> What is to be the nature of Funknet?
>>>> Most of us would think of it as an arena in which we can discuss 
>>>> topics
>>>> within the/a functionalist paradigm by swapping ideas in an ongoing
>>>> conversation. However, the polemical "article of faith" abstract Fritz
>>>> Newmeyer has sent round contains no empirical arguments, so it 
>>>> cannot be
>>>> responded to without reading the entire article. The controversial 
>>>> points
>>>> he makes in the article will basically go unanswered in this forum 
>>>> simply
>>>> because the genre of email doesn't permit a paradigm-level 
>>>> response. Only
>>>> another article can respond to an article. So Fritz gets to trash 
>>>> someone
>>>> else's work in public on the basis of a promissory note.
>>>> It would be sad if Funknet became a clearing house for the exchange of
>>>> titles and abstracts rather than the actual discussion of ideas.
>>>> - Paul
>>>>> Dear Funknetters,
>>>>> I think that some of you might be interested in the following 
>>>>> paper of mine:
>>>>> 'What Conversational English Tells Us About the Nature of Grammar'
>>>>> ABSTRACT It has become an article of faith among many functional 
>>>>> and cognitive linguists that the complex abstract structures 
>>>>> posited by generative grammarians are an artifact of 'disembodied 
>>>>> sentences that analysts have made up ad hoc, ... rather than 
>>>>> utterances produced by real
>>>>>  people in real discourse situations' (Michael Tomasello). Their view
>>>>> is that if one focuses on 'naturally occurring discourse', then 
>>>>> grammar
>>>>> will reveal itself to be primarily a matter of memorized formulas and
>>>>> simple constructions. This paper challenges that view. Basing its 
>>>>> claims
>>>>> on a 170MB corpus of conversational English, it argues that the 
>>>>> nature
>>>>> of real discourse reinforces the need for a sophisticated engine for
>>>>> representing and accessing grammatical knowledge. At a more specific
>>>>> level, it challenges Sandra Thompson's claim that evidence from
>>>>> conversation leads to the conclusion that sentential complements 
>>>>> (e.g.,
>>>>> 'you're ready to go' in 'I guess you're ready to go') are not
>>>>> grammatically subordinate.
>>>>> The paper can be accessed at the following url:
>>>>> Best wishes,
>>>>> Fritz
>>>>> Frederick J. Newmeyer Professor Emeritus, University of Washington
>>>>> Adjunct Professor, University of British Columbia and Simon Fraser
>>>>> University [for my postal address, please contact me by e-mail]

More information about the Funknet mailing list