Wednesday, 9 September 2015



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).

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).


  1. Thanks for the very nice discussion, Gillian! Your fears about readjustment rules are shared by many (Dan Siddiqi's dissertation was built on such fears), and by a former me. But I really feel like what I have come to understand about them is what Alec said in re this very topic in a coffee break at Roots: "Why is everyone so skeptical about phonology?"

    What did he mean by that? A phonological system is never, or at least very rarely, a perfect, completely general, one-set-of-rules-for-all-cases kind of system. There are often subsets of the lexicon that obey phonological generalizations that are distinct from those obeyed in other subsets of the lexicon. Phonologists model such behaviors these days as co-phonologies: mutually coexisting yet partially contradictory phonological systems that apply to distinct subparts of the lexicon.

    ...more in next comment...

  2. Imagine the LAD is a pattern-maximizer. Given say, 3+ instances of a pattern, it finds a rule better than a list, if a plausible rule exists that doesn't violate UG. In the domain of phonology, that would mean that there's lots of tiny co-phonologies floating around: A class of nouns subject to a cophonology where vowel raising applies in certain environments (e.g. plural ones), another class of verbs where vowel shortening applies in certain environments (e.g. participial ones). There is a fact of the matter about that: Either they're listed, or they're generated by rule; we don't have to be utter skeptics about the very possibility that such mini-cophonologies exist. It's a question of how many instances of a given pattern are needed to motivate the cophonology, not a question of whether such subpatterns exist. Another example: Palatalization of obstruents at the end of latinate stems suffixed with -ion, -ian, -ial. We don't want to say that's a general rule of English phonology, but it's general enough that it's productive in its sphere. Solution: morphologically restricted cophonology, which you could model with rules if you're a rule kind of person (readjustment rules) or with constraints if you're a constraint kind of person (OT), or however your favorite model works.

    I think the repugnance that we all feel (in my case, felt) at the concept of readjustment rules had a lot to do with the way that they were introduced in the original paper, as segment-rewriting rules. That made the whole formalism seem completely insane. Why, if I could just posit a rule that rewrites 'oo' as 'ee' (using orthography for conveniencere), I could turn ANYthing into ANYthing, and we'd have no idea of what was a legitimate phonological rule and what wasn't! But that was the most absurd, least charitable interpretation of the presentation of the idea. What seems like a legitimate phonological rule to a theorist, or more to the point to an acquiring child's LAD, will still limit the range of readjustment rules that are possible; it's an empirical question, not an a priori one. The segment-rewrite formalism was just shorthand for e.g. a featural change formalism, where it is much clearer that not anything goes--it might be reasonable to turn a [+obs] segment into a [+palatal] one in the context of a given morpheme, but it's clearly not reasonable to turn [+labial, -stop, -obs] into [+velar, +stop, +voiced], especially on the basis of e.g. just one occurrence; the generalization of a pattern still takes multiple occurrences to motivate, and a certain amount of similarity between variants will need to be preserved to make the rule a better way of storing the information than just listing (the less similarity, the more rules needed to get from a) to b)).

  3. The moral of the story is really that the existence of readjustment rules/mini-co-phonologies is an empirical question. It's an idea that seems to do some useful theoretical work -- if the pie is cut THIS way, then we get *ABA, which we predict for independent theoretical reasons -- and the whole picture hangs together just that much more tightly.

    To be, no doubt, continued. :)

    1. Yes, I totally agree that LAD is a pattern seeking device. And it is an empirical question how many instances you need before you start positing a rule, as opposed to listing. I guess Charles Yang has an answer to this question, or at least a proposal. This to me is an absolutely URGENT question of language acquisition that we could actually potentially solve with careful study, possibly with artificial language learning. My own feeling is that we are not studying this quesition empirically, by and large, but going by common sense assumptions. My feeling is that our gut feeling as linguists is always over eager with the Rule strategy, for perhaps understandable reasons. We have a hammer so everything is looking like a nail. We don't like listing because it takes the phenomenon out of our particular domain of interest and study. Bottom line is that we cannot simply trust our instincts or analytic scientist best judgement here. Ditto for what counts as a `reasonable' phonological rule. (I guess the two questions are intertwined). I always end up disagreeing with my morphology friends about where the cut-off for positing a rule is. So clearly the judgement call varies between individuals and theoretical proclivities. For this reason, we desperately need a strategy for figuring out the cognitive reality behind listing and pattern generalization. This is probably going to involve doing some serious psycho/neuro-linguistics.

      But in addition, even if the worry about readjustment rules goes away because we only posit real, psychologically plausible ones, how does the insertion mechanism distinguish between a suppletive form and an adjusted form for the purposes of *ABA? I mean won't they all have the same index? Put another way, why should having the same abstract phonological entry be sensitive to syntactic structure inclusion relations? I understood it (I thought) when it was a generalization about the `same' item, but now even suppletive roots can count as the `same' item. Maybe there is something I have misunderstood.

  4. Yes, hallelujah that we all recognize that root suppletion exists. This is well known, I'd say, and always should have been: I cite Veselinova 2006, Corbett 2007, Siddiqi 2009, Bobaljik and Harley 2012, Haugen and Siddiqi 2013b, and Harley to appear in my LI paper from earlier this year (here), where I hammer on the idea a bit as well, with facts from Greek (verbal suppletion triggered by tense/aspect for "eat", "see", "say"). I frankly am amazed at how long it took for it to be widely acknowledged that this idea has no clothes.

    As for the unconstrained power of morphophonological re-adjustment rules, Jonathan himself put it well---to quote from my LI paper:

    As Bobaljik (2012:140) refreshingly puts it, "In theory, there is a sharp division of labor between rules of exponence and readjustment rules. . . . Of course, there is a difficult grey area for the analyst in establishing just where the boundary lies; alternations like many – mo-re could be treated formally as suppletion . . . or as the output of a very powerful readjustment rule, rewriting the syllable rime." (Bobaljik himself eschews readjustment rules entirely in his analyses of comparatives. See also Haugen and Siddiqi 2013a for a critical discussion of the explanatory power of readjustment rules, and compare the unmincing condemnation in Bermudez-Otero 2013:83: "DM [Distributed Morphology] routinely . . . resort[s] to devices, like the unconstrained use of readjustment rules, that blur the line between allomorphy and phonology, and destroy the empirical content of the theory.’’) Without a criterion for deciding when a morphophonological readjustment rule is involved, and when simple allomorphy, the appeal to unspecified readjustment rules threatens to be no better than Justice Stewart's famous criterion for recognizing pornography ("I know it when I see it")

    1. Glad to hear that Bobaljik eschews readjustment rules entirely in the comparatives thing. I hadn't remembered any but I probably wasn't paying attention to that in particular when I was reading it. It makes me feel better about the result. Actually, I am sure a lot of cool and robust patterns and facts can be seen without recourse to readjustment rules.

      In fact, I would like to see some examples of claims and generalizations that stand the test of time that crucially rely on readjustment rules before they can emerge. If its just a matter of getting the right form to emerge at the end of the day, then the suppletion mechanism is powerful enough to do that. The fact that you can squint hard and see a family resemblance among the suppletive forms just means that history leaves a lot of litter around, and who knows, general phonological resemblance might actually help with lexical access.

    2. Yeah, I feel the same way. There are a number of independent issues here, I think, and the fact that e.g. umlaut and other assimilative rules that were active in earlier stages of the language leave their mark (and trigger occasional analogical imitation) is in principle open to non-synchronic capturing (generalizations stated over the surface forms, etc: not captured by the synchronic internal grammar per se in the form of readjustment rules). Whether the formal, synchronic system has to countenance this kind of rule is the question, and whether such a rule has the same power and provenance as the conditions on allomorphy that we need for the worst case (i.e., suppletion). Jury's out, I think, but given the claims in the lit, we should err on the side of skepticism that two separate mechanisms exist, and that one is subject to strict locality, and the other not (or to some kind of locality defined over different domains...).

  5. PS: Count me among your "avid blog followers", Gillian!

  6. "Without a criterion for deciding when a morphophonological readjustment rule is involved, and when simple allomorphy, the appeal to unspecified readjustment rules threatens to be no better than Justice Stewart's famous criterion for recognizing pornography ("I know it when I see it")"

    Would it help if we said something like, "Readjustment can change at most 3 features on 2 segments" or similar? Surely we can ask phonologists about this; we syntaxy folks don't have to rely on OUR intuitions about what's a plausible candidate for a phonological process. But (common sense argument again) I don't think it's that crazy to say that we recognize that there's a difference in kind between go~went (no segments in common, only one instance of the alternation pattern) and sleep~slep-, leap~lep-, keep~kep- (all but one segment in common, vocalic alternation of a type attested elsewhere in the language), and to hypothesize that the LAD differentiates too.

    As for *ABA -- it's a generalization about the domains that can condition insertion of vocabulary items. The abstract phonological entry that is the input to readjustment IS the vocabulary item; readjustment rules (or mini cophonologies or whatever) apply to the output of vocabulary insertion. That is, the insertion mechanism doesn't see readjustment, that's downstream. The stem [slip] is inserted, e.g., and then morphophological patterns apply to turn it into slEp in the context of [+past]. Suppletion, on the other hand, occurs when [wEn] is inserted at the node 235GO. [wEn] could itself be subject to morphphonological readjustment; that's what Mercedes and I point out for the Hiaki roots in our 'DM Today' paper. But readjustment (which might not be subject to the same locality conditions as insertion), even if it occurs, is not going to turn the inserted form into something utterly different, so it's not going to yield *ABA.

    ...or maybe I'm misunderstanding the question?

    1. Yes, that´s the kind of explanation I needed to understand how the suppletion thing is going to work. If node 2437a can be spelled out via vocab item I OR vocab item II (suppletion), then the constraint *ABA must be a constraint that applies post insertion. The structure pre-insertion cannot tell the difference between vocab item I and vocab item II. I am sure it can be stated technically so it works, but I just think that acknowledging the index which enforces the `sameness' of the two exponents, is in theoretical tension with the generalization that requires suppletive forms to be importantly different when it comes to insertion.

      As for asking the phonologists, I think you would also probably get a different answer depending in which phonologist you ask. We simply cannot decide this a priori or using elegance considerations or intuition (whether that means phonologists' or syntacticians' is immaterial). I think we have to figure out a way of probing what the actual cognitive fact of the matter is.

  7. As I said in my comment to Jason, I would REALLY like to see some examples of claims and generalizations that stand the test of time that crucially rely on readjustment rules before they can emerge. If its just a matter of getting the end form right, then I can probably just throw in a suppletion mechanism and we can agree to differ on that last step without disturbing the rest of the DM architectural assumptions.