Sociological Linguistics

[Larry Trask]

On Sat, 22 May 1999, Patrick C. Ryan wrote:

[on my denial of his claim that increase in complexity is everywhere the
norm, and on my example of dedicated parasites]

> And is that parasite one-celled? Was it there at the beginning of
> life on earth?

No, and no.  I am speaking of parasites which are descended from
free-living ancestors and which have lost all the organs, structures and
functions not required in becoming specialized parasites.

Stephen Jay Gould's popular book Full House (a.k.a. Life's Grandeur) is
a splendid account of the idea that the history of life on earth is the
history of a monotonic increase in complexity.  Gould is able to show
that this view is a mirage.

> Degeneration of certain functions is always possible. Take Chinese
> as an example. The language from which it is derived did have a
> plural.

Quite possibly, but the loss of a former grammatical distinction cannot
reasonably be described as "degeneration": the term is emotive,
inaccurate and entirely out of place.

Modern English is not more "degenerate" than Old English because it has
lost so much of the earlier morphology: it is merely different.  Nor is
it more ambiguous, nor is it less complex overall: it is less complex in
some particular respects but more complex in others, and no metric
exists for weighing up overall complexity.

[on my point that modern IE languages are no better or worse than PIE
for having generally simpler inflectional morphology]

> It always gets back to "better/worse", "inferior/superior" for the
> sociologically oriented. For them, different is a pejorative. Perhaps
> ambiguity is better, quien sabe?

It wasn't me who introduced value judgements into historical change or
into typology: it was you.

And, no, ambiguity is not the point either.  Daughter languages are not
more or less ambiguous than their ancestors.  Once again, I think you
are confusing grammaticalization with expressive power.  Every language
can make all the distinctions of time its speakers want to make.
Whether a given language does, or does not, build some of those
distinctions into its grammatical system in the form of tense-contrasts
is irrelevant to expressive power.

[on my point that we have no access to the remote antecedents of human
language]

> Yes, Larry, we know this is your belief. But, although one cannot
> convince a believer, let me mention a few things for those readers
> who still have open minds:

> 1) All present languages potentially provide "data" for the
> languages of our earliest ancestors if monogenesis is accepted;

I can't see that it makes the slightest difference whether one embraces
monogenesis, rejects it, or remains agnostic on the issue.  Anyway, the
point is that modern, attested and reconstructed languages are so remote
in time from the earliest beginnings of language that they are incapable
of shedding any light on those beginnings.

> one can study IE, can one not? even though we have no attestation of
> it beyond what we can reconstruct;

I think you mean `PIE'.  Sure, we can study PIE, because we have enough
data from the daughter languages to do so.  But PIE was spoken only
something like 6000 years ago -- a tiny slice of the total time that
human beings have used language.  And we are exceptionally fortunate in
the material available to us in this case.  A very few other languages
-- such as Proto-Oto-Manguean -- have been reconstructed back to around
the same time depth.  But cases involving significantly greater time
depth -- such as the widely accepted Afro-Asiatic family -- have proved
refractory for reconstruction.

> 2) Let us not say that "polygenesis" is against common sense; let us
> rather say that polygenesis is intellectually on a par with
> _believing_ that the earth was created in six days;

Not so, I'm afraid.  The evidence at our disposal at present does not in
any way allow us to pronounce in favor of either monogenesis or
polygenesis.  Most of us, I suppose, lean toward monogenesis as the
simpler scenario, but at present we have no substantial piece of
evidence telling one way or the other.

> 3) I am not the only person who believes that languages of much
> greater antiquity than that allowed by the flawed mathematics of
> glottochronology and lexicostatistics --- linguists, not amateurs
> like myself.

Sorry; this passage has come out garbled into incomprehensibility.
I will hazard the guess that you are proposing that reconstruction is
possible much further back into the past than has so far been achieved.
True, a few linguists -- though not many -- believe this.  Maybe they're
right, but the only way they can persuade the rest of us is by
exhibiting an undeniable instance of it.  And so far that hasn't
happened.

[on my example of dual number]

> If I had meant to include "dual" as a category, the absence of which
> introduces ambiguity, might I not have written that --- just another
> straw-man argument from the master debater.

Eh?  Lacking or losing an overt plural constitutes ambiguity, while
lacking or losing an overt dual doesn't?  What is this supposed to mean?

[on my example of grammaticalized visibility in Kwakiutl]

> Yes, it makes Kakiutl more expressive.

No, it doesn't.  A speaker of English is perfectly able to express
visibility or invisibility on any occasion on which it appears to be
relevant.  A speaker of any language can do the same.  Kwakiutl merely
happens to build this information into its grammatical system, so that
it must be expressed on every occasion, whether this information is
relevant or not.  In the same way, English builds a modest
time-distinction into its grammar, so that this information must be
expressed even when strictly unnecessary, as in my example `I saw Susie
yesterday', in which the tense marking is entirely redundant.

> Now one can say in English: 'I saw Susie'. What is the equivalent
> literal translation in Mandarin?

Depends on the context.  Utterances are not made in a vacuum: they are
made in a context of what is already known or believed and of what has
already been said.  Even in English, I can hardly say `I saw Susie' if I
have no reason to believe that you are acquainted with Susie, or that
you can immediately tell which one I mean of several people called
`Susie'.  Even if I am sure that you and I both know exactly one person
called `Susie', saying `I saw Susie' would be a very odd way of starting
a conversation in most contexts.

[on my statement that English has only a single past tense]

> You may repeat that 'English has only a single past tense' until the
> Fenris wolf swallows the sun, and you will convince only yourself
> and your friends who hold it dogmatically.

This is mysterious.  English has only two tenses: the past tense and the
other one.  I know of no linguist who questions this, unless you count a
few people who use the term `tense' in an eccentric way, like some of
the Hallidayans.

Of course, in addition to its simple tense system, English has a good
deal of aspect, mood and modality built into its verbal system, or at
least into its sentence structure.  Some of these further distinctions
are expressed morphologically, far more of them lexically, and some of
the distinctions are more strongly grammaticalized than others.  The
result is decidedly elaborate and complex, as anyone knows who has tried
to teach English to foreign learners.  But English has only two tenses.
This statement is no more in doubt than the statement that English has
no grammatical gender.

[on the eleven tenses of Bamileke-Dschang]

> B-D is, in this regard, more expressive than English; and also less
> ambiguous.

Nope -- not at all.  I can say `I saw Susie this morning' or `I saw
Susie two months ago', using the same tense-form, and there is no
ambiguity.  Chinese can do the same using no tense-forms at all.  B-D
must use two different tense-forms in the two cases.  But all three are
equally expressive, and equally lacking in ambiguity.

[on PR's accusation that I am doing "feel-good sociology", whatever that
is]

> Your ignorance of sociology is probably an important component of
> why you are unable to distinguish between scientific beliefs
> justified by reasoning from data, and your "positions".

Very droll, no doubt, but completely wide of the mark.  I would like to
say that this is a case of the pot calling the kettle black, but it
seems to be more a case of the pot calling the igloo black.

Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larryt at cogs.susx.ac.uk