Language
Tuesday 23 August 2016
Philosophy, Psychology and Monkey Cognition
Hey there´s this thing called the
European Society of Philosophy and Psychology which has a conference every year. The meeting this year was at the University of St Andrews August 10-13, 2016 (see http://espp16.wp.st-andrews.ac.uk for the conference website).
In fact,
I happen to be the Linguistics Program chair for this society/conference and
so, for my sins, I have to read abstracts, think of keynote speakers and
actually attend the conference every year.
Every year, I wrench myself away from my linguistics-internal concerns
and duties wondering whether I really have time for this. And every year I come
away from the conference thinking "OMG, I am so glad I made time for
this”. We linguists should make time to engage with our
colleagues over in philosophy and psychology who are thinking about the very same issues but in radically different and yes, sometimes
incompatible ways. We can make all the
noises we like about building bridges with philosophy and psychology in the
abstract, but unless we talk to them and go to their conferences those bridges
won´t actually get built, and misunderstandings will proliferate. In particular, clicking on and reading the
occasional hyped psychology article that catches your eye does not prepare you
for the whole culture of concerns and assumptions and indeed heterogeneity of
approach that you find when you are actually at one of these meetings.
This
year, there was a common thread running through the conference on animal
cognition and primate cognition in particular, in part due to the invitation of Philippe
Schlenker, CNRS-Institut Jean Nicod (http://www.institutnicod.org/membres/membres-permanents/schlenker-philippe/?lang=en ), as one of the keynote speakers and to the
local expertise of our hosts at St Andrews in the area of primate cognition
(Check out the Centre for Social Learning and Cognitive Evolution here https://risweb.st-andrews.ac.uk/portal/en/organisations/centre-for-social-learning--cognitive-evolution(aca65ea5-18be-4425-8477-a13cdbd890c9).html )
As we all probably know, the field of primate cognition today is largely
insensitive to the overly simplistic historical dichotomies of `innate´ vs.
`learned´ that inspired linguists´ early
attempts to teach chimpanzees human language (and which still fills chapters in
beginner textbooks on psychology of language).
A linguist looking to sign up for the cheering gallery on one side or
other of that debate will be cruelly disappointed. So its all the more important for the responsible linguist to get up to speed on the latest knowledge
that has emerged from research in this area over the past couple of decades.
First of
all, it is quite convincing from the research that the great apes communicate
intentionally, do social learning, use tools, have problem-solving abilities
and even some hallmarks of theory of mind.
The whole issue of intention to communicate, as tested and
proved by a number of researchers, requires at least some sort of
recognition that others possess minds, and that the state of their knowledge
can be affected by one´s own communicative actions (I am thinking here of
papers I heard by Christine Sievers (https://philsem.unibas.ch/seminar/personen/sievers) and Thibaud Gruber (https://www2.unine.ch/compcog/thibaud_gruber) , and also Katie Slocombe´s (https://www.york.ac.uk/psychology/staff/faculty/ks553/) contribution in the invited
symposium). So the interesting question
is whether there is some sort of basic groundfloor `theory of mind´ such that
apes and very young children can have that,
but not the fullblown version that would
allow them to pass the standard false belief test. If there is an intermediate version, is it
distinguished from the full version because:
-It ascribes
knowledge to other minds rather than
belief (as Jennifer Nagel was arguing http://individual.utoronto.ca/jnagel/Home_Page.html
)
-It is
expressive rather than genuinely perspective shifting (as Dorit Bar-On
suggested http://www.doritbar-on.com )?;
-It is
one-step rather than recursively specular in the way it takes other minds into
account?
Are any
of these distinctions themselves correlated with any of the others, or indeed
with the the special linguistic capacities for syntax?
In
Schlenker’s keynote address that kicked off the conference, he described in
detail the sign system employed by a number of groups of monkeys and attempted
to describe the truth conditions for each individual sign in a systematic
way. Schlenker was scrupulous in staying
away from questions of whether we should call such systems `language’ or not, preferring rather be specific about
how it seems that this particular system is working. The interesting aspect of his proposal for
the meanings deployed in the system is that they seem to involve a kind of
blocking, where the informationally more specific sign blocks the use of the
more general one. If this kind of
informational pragmatic choice guides the deployment of signs in monkey
communities, it seems to point to at a least a limited theory of mind.
Schlenker himself, in ascribing pragmatic competence to monkey groups, stopped
short of claiming that they needed to possess full specular theory of mind of
the kind that is assumed to lie behind our competences in standard forms of
Gricean reasoning. But if this limited
kind of pragmatic effect can be seen even in these simple systems, with
rudimentary acknowledgement of audience then it is important both for studies
of the evolution of theory of mind, and for the architecture of the language
faculty. (If you are interested in the details, see the recent issue of
Theoretical Linguistics devoted to Formal Monkey Semantics for a target article
written by Schlenker and his research group, with commentaries by various
linguists and cognitive scientists http://www.degruyter.com/view/j/thli.2016.42.issue-1-2/issue-files/thli.2016.42.issue-1-2.xml
)
What does
it mean that monkeys seem to have systems that are best described via
informational blocking (maxim of quantity).
Full Gricean reasoning probably involves infinite specular regress, but
certain facts with respect to informational computation do not seem to require
that degree of sophistication. It can
be shown that great apes exist in social groups and utilize their gestures and
vocalizations intentionally with an aim to express or convey information. But surely one can intend to warn others without having
full blown theory of mind? Some acts
of expression can be purely reflexive in the sense of not being under conscious
control, but it can be shown that at least some ape gestures/vocalizations are
not of this type. And pragmatic effects
can only arise in the context of an audience. So there must be different levels of
pragmatics corresponding to different levels of sophistication with respect to how we represent other minds.
Andrew
Whiten from St Andrews gave the final keynote of the conference and he talked
about the work that his group has been doing on cultural transmission within
primate communities. With respect to
being social animals with socially transmitted traditions, chimps once again
seem to have some restricted version of what we see in humans. In fact, Whiten would argue that the
differences we find here are slight indeed.
Chimps use tools and pass on use of those tools, as well as certain
non-necessary ways of doing things to the group. Groups of chimps also seem to have
a strong instinct for social conformity within the group. Chimps are good at imitating and are rational
problem solvers/learners. Unlike the
claims of the Tomasello group, the Whiten group in St Andrews has been
successful in showing that chimps do have imitative and social instincts---
they are not just emulators. (Tomasello’s group had a hypothesis to the effect
that chimps merely tried to emulate `goals’ of
actions they perceive others doing where that goal is attractive to them, but
are sloppy about the detail when copying the means by which the emulated being
is bringing about those goals. So for the Tomasello group apes were not genuine imitators in our sense.)
In fact,
it can be shown that these cousins of ours, the great apes, are actually
qualitatively better imitators than other monkeys. (Ironically then, there is actually no generalized `monkey
see, monkey do’, but a restricted version of that is found specifically within the
great apes.) Chimps can be taught a version of the Simon Says game (for a
reward) very quickly for example, but not capuchin monkeys who are otherwise
pretty smart. But the
Tomasello group is partly right too. In
experiments where a certain goal is achieved by a sequence of actions, a chimp
will copy the actions to achieve the goal. But if it becomes manifest to the
chimp that some of the steps are not practically necessary to the observed
outcome, the chimp will miss out these steps.
Young children however, will systematically continue to repeat the
useless elaborative steps, even after it has been made manifest that they do
not contribute to the outcome. This the
phenomenon that is now known as `over-imitation’. They were initially
discovered by the Whiten group, and have been robustly replicated.
But maybe
children are less rational at this age than the chimps, or less able to
calculate physical outcomes, so they are just playing it safe? Or maybe
both children and human adults in an experimental situation see the task as a
kind of game which leads them to over imitate?
In an interesting new extension of the paradigm, the Whiten group ran a
similar experiment with adults in a non-experimental situation. They set up the same primate-tested tasks as
part of a hands-on installation at Edinburgh zoo, inviting adults passing
through the exhibit to `have a go’ at
the tasks that had been tested on primates. The adults could watch the training
video (the different conditions were cycled) and then attempt the same task on
the actual equipment. Importantly, the
human adults did not know they were even being observed, although they were in
fact being filmed. Once they did the
task, an experimenter came up to them and explained `candid camera’ style, asking if they were willing to sign a
consent form for their data to be used (they usually did). The amazing upshot of the study was that the
unobserved fully rational adults also did the full detail imitation even
when they could judge that those extra bells and whistles had no effect on the
outcome! Recall that the chimps in the same situation left out the extra
bits and went straight for the prize.
Now who
would have thought that there would be so much interesting difference in the
realm of imitation? Imitation, we are
told in LING 101 is the thing that
language acquisition is not about.
For good reason, since imitation alone is totally inadequate to the
task. In particular, overexuberant imitation of everything would be a
hopelessly huge task and would actually inhibit pattern discovery (in the
jargon, compulsive imitation is not the same as over-imitation). There needs to be selective attendance to
certain aspects of what the young human is exposed to. But that is not all, I would argue. Over and above that, the evidence seems to be that humans in certain domains attend
and imitate in an overly fine-grained way, in a way that does not need to be justified by immediate practical goals.
Another
difference in the imitative capacity between us and chimps, is called
`ratcheting’, and I think it actually might be related to the first. Children can easily be taught to build one
learned behaviour on top of another one.
If you try to do this with chimps, they get stuck at the first
stage. Now, if you try to teach them the same
sequence of actions but to a single final goal, they are capable of that, so its not the memory or extended nature of
the task that is hard. Things go wrong
if you teach a chimp a behaviour that achieves a certain goal, and then, a
while after they have learned that behaviour, try to teach them to suspend goal
number one and use the first behaviour as a stepping stone add a new behaviour and achieve an even
bigger payoff. Chimps should in
principle be rational enough to see the advantage of this but in fact,
according to Whiten, they get stuck.
No ratcheting (this is
apparently the technical term but I might be spelling it wrong). This
for the Whiten team is the reason that chimp culture does not undergo
cumulative advance, unlike our own. But
the discussion of the no ratcheting discovery started me thinking about how
this could also be related to the acquisition of language.
Rational
and goal oriented emulation like the chimps generally do ends up as low
fidelity copying--- you concentrate on the outcome and try to reproduce that.
But arbitrary and causally more opaque tasks (or tasks with non transparent or
deferred payoffs) are hard to
acquire if you are a goal-oriented learner.
But what if you just take pleasure from your
success at high fidelity imitation with no need of reward? What if we humans have an instinct for
learning just for the fun of it (within certain targeted domains of course, that
we are predisposed to pay attention to) ?
Learning for the sake of it, non-goal
oriented learning, is actually something that distinguishes us from our
great ape cousins. That would explain
both the rachetting and the overimitation.
Learning is not a rational strategy for us the way it is for chimps. We
get rewards just from the success of learning itself. Arguably both rachetting
and overimitation are important for learning language. Overimitation because
very fine grained motor imitations are necessary to start producing the differentiation
in the produced code that language requires, even in the absence of immediate
rewards. Ratcheting because we need to be able to build up our skills
cumulatively and build hierarchical complexity in the symbolic system. Recursion at the symbolic level requires that
one can use one result as a stepping stone to the next. The very thing that the chimps get stuck on.
(Now there may be evidence that chimps use recursive reasoning in problem
solving tasks, or in their vision systems like us, or whatever, but what we are
talking about here is symbolic manipulation, which is crucially mediated by learning.) So one hypothesis might be that it is
narrowly goal-oriented learning vs. and instinctive joy of imitation that is
one of the crucial ingredients in what makes us special. (This latter aspect on
the other hand, we seem to share with some birds).
Back to
Tomasello. Tomasello’s big idea about what makes humans unique is the complexity
and richness of our social structures (http://www.eva.mpg.de/psycho/staff/tomas/pdf/Tomasello_EJSP_2014.pdf ).
However,
my own hunch from the research of the Whiten group and others is that what we are seeing in chimps with respect to social structures is a difference in degree not in kind, and certainly not drastic enough to underpin the huge cognitive leap to language. If so, then we need to look elsewhere for the crucial cognitive
ingredient in my opinion. One could speculate about a kind of genetic switch
that suddenly allowed recursion. But what would that be? Suppose the key is in
the non-goal orientedness of the learning mechanism, which creates both high
fidelity and ratcheting?
Thinking
about it this way makes a nonsense of the old dichotomies of nature vs.
nurture by the way--- the thing that is innate and
distinctive is a way of learning. (The
more we know about nature and nurture from the geneticists anyway the more
those two things get blurred.)
At any
rate, that is just a flavour of the ideas floating around the conference and my
own thoughts on hearing them. To be clear, none of the speculations and
ramblings opined in this piece would be endorsed by those real psychologists
and philosophers out there, but they certainly provoked me to think them. I can only hope my short description inspires
other linguists to attend this kind of meeting in future. The next ESPP takes place at the University
of Hertfordshire. Watch this space !
Monday 14 September 2015
Allosemy---- No thanks.
On Allosemy
It seems like I am always complaining about the status of
semantics in the theory of grammar. I complain when its ignored, and then I
complain when its done in a way I don´t like, I complain and complain. Today is not going to be any different.
At the ROOTS IV conference, we had a number of lexical
semantics talks, which clearly engaged with meaning and generalizations about
root meaning. Then we had the morphology talks. But I´m not convinced those two groups of
people were actually talking to each other.
Now, the thing about Distributed Morphology is that it doesn’t believe
in a generative lexicon, so all of the meaning generalizations that are in the
lexicon for the lexical semanticists have to be recouped (if at all) in the functional structure, for DM and its
fellow travellers, me included. This is not a deep problem if we are focusing
on the job figuring out what the meaning
generalizations actually are in the first place, which seems independent of
arguing about the architecture. But there is also a danger that the
generalizations that the lexical semanticists are concerned about are perceived
as orthogonal to the system of sentence construction that morphosyntactians are looking at. Within DM, the separation of the system into
ROOT and functional structure already creates a sharp division whereby meaty
conceptual content and grammatically relevant meanings are separated
derivationally. This in turn can lead to
a tendency to ignore lexical conceptual semantics if you are interested in
functional morphemes, and to suspect that the generalizations of the lexical
semanticists are simply not relevant to your life (i.e. that they are not part
of the `generative system´). To the
extent that there are generalizations and patterns that need to be accounted
for, we need to look to the system of functional heads proposed to sit above
the verbal root in the little vP. But
more challengingly, we need to relate them via selectional frames to the sorts
of ROOTS they combine with in a non ad hoc manner. If, in addition, we require a constrained
theory of polysemy, the problem becomes even more complex. I think we are nowhere close to being able to
solve these problems. Perhaps because of
this, I think that standard morphological and syntactic theories currently do not yet
engage properly with the patterns in verb meaning, by which I mean both
constraints on possible meanings, and the existence of constrained
polysemies. I contend that the architecture
that strictly separates the conceptual content of the root from the functional
structure in a derivational system must resort to crude templatic descriptive
stipulations with which to handle selection.
This architecture also obscures the generalizations surrounding
polysemy.
One of the interesting talks in the conference that was one
of the few that attempted to integrated worries about meaning into a system
with DM-like assumptions, was the contribution by Neil Myler. Neil was
interested in tackling the fact that the verb have in English is found in a wide variety of different
constructions, and he was interested in giving a unified explanation of that
basic phenomenon. To that extent, I
thought Neil´s contribution was excellent, and I agreed with the motivation,
but I found myself uncomfortable with
some of the particular tools he used to put his story for have together. The issue in question involves the deployment
of Allosemy.
Let me first complain about the word Allosemy. It´s
pronounced aLOSSemi, right? That´s how
we are supposed to pronounce it. Of course, doing so basically destroys all
recognition of the morphemes that go into making it , and renders the word
itself semantically opaque even though it is perfectly compositional.
I hate it when stress shift does that.
Curiously, the problem with the pronunciation is similar to
the problem I have with its existence in
the theory, namely that it actually obscures the semantics of what is going on,
if we are not careful with it.
Let´s have a look at how Allosemy is deployed in a series of recent works by Jim Wood, Alec
Marantz and Neil Myler (We could maybe call them The NYU Constructivists for short). I am supposed to be a fellow
traveller with this work, but then why do I feel like I want to reject most of
what they are saying ?? Consider the
recent paper by Jim Wood and Alec Marantz, which you can read here .
So to summarize briefly, the idea seems to be that instead
of endowing functional heads with a semantics that has to remain constant
across all its instantiations, we give a
particular functional head like little v N possible semantic meanings, and then say
that it is allosemic. In other words it is N-ways ambiguous
depending on the context. This allows
syntax to be pure and autonomous. As a
side effect this means that meaning can be potentially built up in different ways, and the same structure can
have different meanings. The cost?
COST 1: In
addition to all the other listed frames
for selection and allomorphy, we now have to list for every item a subcategorization
frame that determines the allosemic variants of the functional items in the
context of insertion. (Well, if you like construction
grammar……)
COST 2: Since the mapping between syntactic structure
and meaning can no longer be relied upon, there is no chance of semantic and
syntactic bootstrapping for the poor infant trying to learn their
language. I personally do not see how
acquisition gets off the ground without bootstrapping of this kind.
COST 3: (This is
the killer). Generalizations about hierarchy and meaning correspondences like
the (I think exceptionless) one that syntactic embedding never inverts
causational structure is completely mysterious and cannot fall out naturally
from such a system (see this paper of mine for discussion).
PAYOFF: Syntax gets to be autonomous again.
But wait. We want this exactly, Why? Because Chomsky showed us the generative
semanticists were wrong back in the sixties?
And anyway, isn’t
syntax supposed to be quite small and minimal now, with a lot of the richness
and structure coming from the constraints at the interface with other aspects
of cognition? Doesn’t this lead us to expect that abstract syntactic structures
are interpreted in universally reliable ways?
Allosemy says that the only generalities are syntactic ones.
Like `I have an EPP feature’ or` I introduce an argument’. It denies that there
are any generalities at the level of abstract semantics. I would argue rather that the challenge is to
give these heads a general enough and underspecified
semantics so that the normal
compositional interaction with the rest of the structure these things compose
with will give rise to the different polysemies seen on the surface. Allosemy is not the same as compositionally potent underspecification. The
strategy of the Woods and Marantz paper is to go for a brute force semantic
ambiguity which is controlled by listing selectional combinations. It is perfectly clear that this architecture
can describe anything it wants to. And while one might be able to do it in a
careful and sensible way so as to pave the way for explanation later on, it is
also perfectly clear that this particular analytic tool allows you to describe
loads of things that don’t actually exist!
So, isn’t this going backwards, retreating from explanatory adequacy?
Of course, the rhetoric of the Woods and Marantz paper
sounds lovely and high-minded. The head that introduces arguments (i* ) is
abstract and underspecified. The kind
of thing a syntactician can love. (There
is also another version of i* which is modulated by the fact that a ROOT is adjoined to it, and this version is
the one that introduces adjuncts and is influenced by the semantics of the ROOT
that adjoins to it). However, core i* is
nothing nothing new, in fact it is a blast from the past (not in a bad way, in
fact). It is just a notational variant of
the original classical idea of specifier, where it was the locus for the
subject of predication (as in the the classic and insightful paper by Tim
Stowell from 1982, Subjects across Categories here). And the i* with stuff adjoined to it is what
happens when you have an argument introduced by a preposition. So i* is only
needed now because we got rid of specifiers and the generality of what it means
to be a specifier.
So. Allosemy. Can we just not do this?
Wednesday 9 September 2015
THOUGHTS ON ROOTS IV, NYU
THOUGHTS AFTER ROOTS
IV, NYU
It’s been a while since New York, but I whisked away for
vacation time immediately afterwards, from which I am only slowly recovering. Many of you will already know that I am also
on sabbatical this term, hanging out in Edinburgh, loosely affiliated with the
University, but trying to lay low. This has in turn made August a month of moving and organizational
hecticness. But productivity is slowly
picking up.
ROOTS IV took place
in New York, June 29th- July 2nd, the 4th meeting of its kind, organized brilliantly
by Itamar Kastnar, Alec Marantz and the department at NYU and co-sponsored by
NYU Abu Dhabi. Check out the website for the panel discussion here,
including a YouTube video of all the panel presentations, including yours truly here.
Avid blog followers will recall that I expressed my fears in
advance of this meeting that I might end up at the wrong party, i.e. that the
workshop would largely be some kind of theory-internal Distributed Morphology
discussion. Alec debunked that notion
forcefully and convincingly in his opening address. And indeed, one can see
from the invited participants to this event, that we were not all specifically classic
DM-ers, but came from a broad group made
up of what Alec called `fellow-travellers’.
By this I think he meant those who broadly shared enough starting
assumptions to actually get a meaningful and stimulating conversation going
about details. As a fellow-traveller, I
offer some thoughts in this blog inspired and stimulated by being at this
workshop and being part of the ROOTS IV event. In the end, the conference split quite firmly
into the morphologists (that group of fellow travellers) and the lexical
semanticists who didn’t actually seem to be in the same conversation (but more
about this in the next post).
MAJOR NEWS FLASH (FOR
ME, ANYWAY)!
It seems to me that at this conference, Distributed
Morphology officially acknowledged in a common and public forum that root suppletion exists. Heidi Harley’s
poster child case from root suppletion in Hiaki has stood up to scrutiny and we
have to just suck it up.
The DM-ers at the conference seemed to all reluctantly agree, including Alec (Skepticism and vocal disagreement from Hagit Borer notwithstanding).
Since it is a little outside my world view, I took some time
to reflect on the special status of roots within DM and what work it does in
the theory. In DM, recall, Roots are the
only listed thing there at the start of
the syntactic derivation. Unlike
vocabulary items, they are not late-inserted. They also have no syntactic features on them
inherently, and they usually come in at the lowest part of the tree (more recent approaches also allow roots to
be `adjoined’ to various syntactic heads,
but we put this aside for now). Roots
are the creatures that anchor the whole derivation, within the theory of Distributed Morphology, and which are the
basis for the enclosing identity within which competition for insertion can be
calculated. They are also the identity
that underlies allomorphy and allosemy in particular contexts. What the fact of root suppletion does to this
system is that, previously, an abstract
phonological representation could be thought to be a stand-in for the identity
represented by a particular root. But if
there is root suppletion then that is no longer always the case, and the thing
that is the same across all spell-outs of ROOTs in a context has to be much
more abstract than that (Heidi makes this point in the article I linked to
above. In that work, she argues for a system of abstract indices to track the
identities we need). I guess this is also the reason that the paradigm people believe
in paradigms. Paradigms are probably a notational variant of the abstract
indices idea (a sub-list defined by
features inside a single address).
To see how this affects the whole system, consider the nice
*ABA generalization that Jonathan Bobaljik has famously proposed and discussed
in his book on comparatives. (Norbert
discussed this work warmly in his blog earlier this summer here ).
*ABA is a constraint
that makes references to a particular kind of situation where syntactic
features are in a particular inclusion relation, ordered in a particular
hierarchy. In this situation, if you have a vocabulary item that can spell out
a lower position but a suppletive one that spells out an intermediate position,
then you cannot revert to the first item to spell out the highest node. Thus the claim is about the correlation between
possible polysemies and syntactic structures—polysemy must respect the contiguity
of the inclusion relations in syntactic
structure, as a constraint on the operation of the Elsewhere Principle. A very
interesting proposal, if true. Now, what
we need to understand about this pattern is that the statement of it also
relies on correctly distinguishing cases of true suppletion from other kinds of
phonological variants in the vocabulary items.
We all understand and accept cases of phonologically conditioned
allomorphy, where the phonological rules present and active in the language
create variations on the ROOT’s abstract representation due to phonological
context. But there are also cases of phonological readjustment
rules that exist in DM, which are sensitive to morphosyntactic context (not
phonology), and which are not the same as
any actual phonological rule in the
language, ( or even possible rule sometimes).
These abstract readjustment rules do not count as suppletion--- crucially
do not `count’ as creating a B out of an
A. Essentially, you still have an A if
you `phonologically readjust’. There
are many of us who do not like ad hoc phonological readjustment rules, just to
preserve the fiction of phonological ROOT identity. But according to Bobalijk (pc), readjustment
rules were crucially taken into account in reaching the *ABA generalization in the first place. (Thanks to Peter Svenonius for pointing this
out to me). Putting this together with
the previous point, consider now the fact that
root identity is no longer underwritten always by an abstract phonological representation, but
by something MUCH more abstract, like an index.
Now we need to make sure that we have an architecture of the kind that constructs ROOT identity across
suppletive environments, while still maintaining an internal distinction
between `related’ variants and
suppletive variants of the same thing for the purpose of stating the deep
Bobaljik generalization. So what gives?
Are suppletive variants `the same’? Or
are they `different’ , i.e. Bs as opposed to As in Bobaljik’s
generalization?
I for one would like to give up ad hoc phonological
readjustment rules in favour of straight-up variant insertion, making these
kinds of variations indistinguishable from cases of suppletion (which we can no longer run away from theoretically, if Heidi is right). But then I am in danger of losing *ABA. Or rather, I would have to make
*ABA a bit of telling historical detritus, a morphological patterning that
shows us something real, but indirectly
and not synchronically. I would also
expect in that case to see some evidence of pure
*ABA where one only needs to compare two distinct forms without the help
of phonological readjustment rules. I
don’t control the examples from the book well enough to know how much reliance
there is on those in Bobalijk’s book to make the generalization.
But in any case, there is a real tension here I think. If there really is a generalization
concerning the mapping between insertion and syntactic structure that relies on
suppletive forms being different in an important sense, then how does that
reconcile with ROOTs having an identity across suppletive variants?
Morphologists: Help?
This has gone on too long.
In my next post on ROOTS IV, I will muse on semantics and the existence
of Allosemy (or not).
Thursday 18 June 2015
Anticipation: Roots
ROOTS
The recent meeting of syntacticians in Athens has whet my appetite for big gatherings with lots of extremely intelligent linguists thinking about the same topic, because it was so much fun.
At the same time, it has also raised the bar for what I think we should hope to accomplish with such big workshops. I have become more focused and critical about what the field should be doing within its ranks as well as with respect to communication with the external sphere(s).
The workshop I am about to attend on Roots (the fourth such) to be held in New York from June 29th to July 3rd, offers a glittering array of participants (see the preliminary program here http://wp.nyu.edu/roots4/wp-content/uploads/sites/1403/2015/02/roots4_program.pdf ), organized by Alec Marantz and the team at NYU.
Not all the participants share a Distributed Morphology (DM)-like view of `roots’, but all are broadly engaged in the same kinds of research questions and share a generative approach to language. The programme also includes a public forum panel discussion to present and discuss ideas that should be more accessible to the interested general public. So Roots will be an experiment in having the internal conversation as well as the external conversation.
One of the things I tend to like to do is fret about the worst case scenario. This way I cannot be disappointed. What do I think is at stake here, and what is there to fret over in advance you ask? Morphosyntax is in great shape, right?
Are we going to communicate about the real questions, or will everyone talk about their own way of looking at things and simply talk past one another? Or will we bicker about small implementational issues such as should roots be acategorial or not? Should there be a rich generative lexicon or not? Are these in fact, as I suspect, matters of implementation, or are they substantive matters that make actual different predictions? I need a mathematical linguist to help me out here. But my impression is that you can take any phenomenon that one linguist flaunts as evidence that their framework is best, and with a little motivation, creativity and tweaking here and there, that you can give an analysis in the other framework´s terms as well. Because in the end these analyses are still at the level of higher level descriptions, and it may look a little different but you can still always describe the facts.
DM in particular equips itself with an impressive arsenal of tricks and magicks to get the job done. We have syntactic operations of course, because DM prides itself on being `syntax all the way down´. But in fact, but we also have a host of purely morphological operations to get things in shape for spellout (fission, fusion, impoverishment, lowering what have you), which are not normal actions of syntax and sit purely in the morphological component. Insertion comes next, which is regulated by competition and the elsewhere principle, where the effects of local selectional frames can be felt (contextual allomorphy and subcategorization frames for functional context). After spellout, notice that you still get a chance to fix some stuff that hasn´t come out right so far, namely by using `phonological´ readjustment rules, which don´t exist anywhere else in the language´s natural phonology. And this is all before the actual phonology begins. So sandwiched in between independently understood syntactic processes and independently understood phonological processes, there´s a whole host of operations whose shape and inherent nature look quite unique. And there´s lots of them. So by my reckoning, DM has a separate morphological generative component which is different from the syntactic one. With lots of tools in it.
But I don´t really want to go down that road, because one woman´s Ugly is another woman´s Perfectly Reasonable, and I´m not going to win that battle. I suspect that these frameworks are inter translatable and that we do not have, even in principle, the evidence from within purely syntactic theorising, to choose between them.
However, there might be deep differences when it comes to deciding what operations are within the narrow computation and which ones are properties of the transducer that maps between the computation and the other modules of mind brain. So it´s the substantive question of what that division of labour is, rather than the actual toolbox that I would like to make progress on.
To be concrete, here are some mid-level questions that could come up at the ROOTs meeting.
Mid-Level Questions:
A. Should generative aspects of meaning be represented in the syntax or the lexicon? (DM says syntax)
B. What syntactic information is borne by roots? (DM says none)
C. Should there be late insertion or should lexical items drive projection? (DM says late insertion)
Going down a level, if one accepts a general DM architecture, one needs to ask a whole host of important lower level questions to achieve a proper degree of explicitness:
Low-Level Questions
DM1: What features can syntactic structures bear as the triggers for insertion?
DM2: What is the relationship between functional items and features? If it is not one-to-one, can we put constraints on the number of `flavours` these functional heads can come in?
DM3: What morphological processes manipulate structure prior to insertion, and can any features be added at this stage?
DM4: How is competition regulated?
DM5: What phonological readjustment rules can apply after insertion?
There is some hope that there will be a discussion of the issues represented by A, B and C above. But the meeting may end up concentrating on DM1-5.
Now, my hunch is that in the end, even A vs. B vs. C are all NON-ISSUES. Therefore, we should not waste time and rhetoric trying to convince each other to switch `sides’. Having said that, there is good evidence that we want to be able to walk around a problem and see it from different framework-ian perspectives, so we don’t want homogeneity either. And we do not want an enforced shared vocabulary and set of assumptions. This is because a particular way of framing a general space of linguistic inquiry lends itself to noticing different issues or problems, and to seeing different kinds of solutions. I will argue in my own contribution to this workshop on Day 1, that the analyses that adopt as axiomatic the principle of acategorial roots prejudges and obscures certain real and important issues that are urgent for us to solve. So I think A, B and C need an airing.
If we end up wallowing in DM1-5 the whole time, I am going to go to sleep. And this is not because I don’t appreciate explicitness and algorithmic discipline (as Gereon Mueller was imploring us to get more serious about at the Athens meeting), because I do. I think it is vital to work through the system, especially to to detect when one has smuggled in unarticulated assumptions, and make sure the analysis actually delivers and generates the output it claims to generate. The problem is that I have different answers to B than in the DM framework, so when it comes to the nitty-gritty of DM2,3 and 5 in particular, I often find it frustratingly hard to convert the questions into ones that transcend the implementation. But ok, it’s not all about me.
But here is some stuff that I would actually like to figure out, where I think the question transcends frameworks, although it requires a generative perspective.
A Higher Level Question I Care About:
Question Z. If there is a narrow syntactic computation that manipulates syntactic primes and has a regular relationship to the generation of meaning, what aspects of meaning are strictly a matter of syntactic form, and what aspects of meaning are filled in by more general cognitive processes and representations?
Another way of asking this question is in terms of minimalist theorizing. FLN must generate complex syntactic representations and semantic skeletons that underwrite the productivity of meaning construction in human language. What parts of what we traditionally consider the `meaning of a verb’ are contributed by (i) The narrow syntactic computation itself, (ii) the transducer from FLN to the domain of concepts (iii) conceptual flesh and fluff on the other side of the interface that the verb is conventionally associated with.
Certain aspects of the computational system for a particular language must surely be universal, but perhaps only rather abstract properties of it such as hierarchical structuring and the relationship between embedding and semantic composition. It remains an open question whether the labels of the syntactic primes are universal or language specific, or a combination of the two (as in Wiltschko’s recent proposals). This makes the question concerning the division of labour between the skeleton and the flesh of verbal meaning also a question about the locus of variation. But it also makes the question potentially much more difficult to answer. To answer it we need evidence from many languages, and we need to have diagnostics for which types of meaning we put on which side of the divide. In this discussion, narrow language particular computation does not equate to universal. I think it is important to acknowledge that. So we need to make a distinction between negotiable meaning vs. non-negotiable meaning and be able to apply it more generally. (The DM version of this question would be: what meanings go into the roots and the encyclopedia as opposed to meaning that comes from the functional heads themselves).
There is an important further question lurking in the background to all of this which is of how the mechanisms of storage and computation are configured in the brain, and what the role of the actual lexical item is in that complex architecture. I think we know enough about the underlying patterns of verbal meaning and verbal morphology to start trying to talk to the folks who have done experiments on priming and the timing of lexical access both in isolation and integrated in sentence processing. I would have loved to see some interdisciplinary talks at this workshop, but it doesn’t look like it from the programme.
Still, I am going to be happy if we can start comparing notes and coming up with a consensus on what we can say at this stage about higher level question Z. (If you remember the old Dr Seuss story, Little Cat Z was the one with VOOM, the one who cleaned up the mess).
When it comes to the division of labour between the knowledge store that is represented by knowing the lexical items of one’s language, and the computational system that puts lexical items together, I am not sure we know if we are even asking the question in the right way. What do we know of the psycholinguistics of lexical access and deployment that would bear on our theories? I would like to get more up to date on that. Because the minimalist agenda and the constructivist rhetoric essential force us to ask the higher level question Z, and we are going to need some help from the psycholinguists to answer it. But that perhaps will be a topic for a different workshop.
Subscribe to:
Posts (Atom)