Date: Fri, 26 Apr 2002 14:12:06 +0200
From: Marina Tzakosta <M.Tzakosta@let.leidenuniv.nl>
Subject: McCarthy (2001) A Thematic Guide to Optimality Theory
McCarthy. John J. (2001) A Thematic Guide to Optimality Theory. Cambridge University Press, xiv+317pp, hardback ISBN 0-521-79194-4, $69.95
Marina Tzakosta, University of Leiden Center for Linguistics
The Thematic Guide to Optimality Theory (hereafter TGOT) by John McCarthy is what its author calls it; a guide to OT 'organized thematically, focusing on concepts rather than phenomena' (p. xi). Consequently, the TGOT is not a textbook, that is, it is not the type of book one should start his readings in Phonology and OT with. It is correctly characterized by its author as 'an adjunct to the traditional textbook or ... 'a supplement to materials prepared...' (p. xii). It is a fascinating book written in such a way that on the one hand difficult notions are simplified, but on the other a well-established theoretical background is required from the reader. For that reason I would suggest it to students who already know how things go in Phonology in general and OT in particular and are informed on the ongoing debate about the degree to which OT contributes to theory. It is also intended for researchers who are interested in OT and its implications not only for Phonology but also Morphology, Syntax, Learnability and Language variation/ change. The book is consisted of 4 chapters, each with its conclusions and suggested further readings.
Chapter 1. The Core of Optimality Theory
The first chapter is an introduction to the general premises and the basic architecture of OT. It is divided in two parts; in the first part the fundamental mechanisms, EVAL and GEN, are defined and their operation is outlined. Core notions such as 'constraint typology', 'constraint schemata', 'comparative tableaux', 'local conjunction', 'constraint interaction' and their implications for faithful and unfaithful mappings and blocking effects are also exemplified. In the second part of this chapter the author refers to 'How to do OT'. I find this part absolutely inspiring and original in the sense that its content as such is not found in other OT textbooks. McCarthy very clearly explains the steps that one must take when doing OT, how the known constraints should be ranked, how the relevant candidates must be chosen, which problems one can come across, how they can be overcome by the introduction of new constraints, and how newly proposed constraints must be formulated. To my knowledge, it is the first time after the rise or OT that the criteria according to which new constraints should be proposed are formally recommended. Suggesting 'correct' constraints is something that has very important implications for the work factorial typologies do and the universals that come up from them. The limitations imposed by these criteria actually put an end to the absurd and endless introduction of new constraints in the problematic cases we cannot deal with, something that underestimates OT and makes it look unable to give specific answers to important issues of Universal Grammar.
Chapter 2. The Context of Optimality Theory
Chapter 2 provides the historical and scientific background within the bounds of which OT has developed. The author comments on the basic premises of the models preceding OT, namely SPE (Chomsky and Halle 1968), Natural Phonology (Stampe 1973), Nonlinear Phonology (Goldsmith 1976) and later theories such as the Principles and Parameters model (see Hayes 1980 for a parametric theory of stress) and Harmony Theory (cf.). The author's aim is twofold: (a) to point out the drawbacks of previous approaches and their failure to account for certain phenomena, something which gave ground to the development of OT and (b) to demonstrate that traditional notions such as 'processes', 'well-formedness', 'filters', 'principles', 'conditions' are still valid in OT, though in a different way. In that sense McCarthy restores the misunderstanding regarding many researchers' belief that OT actually seeks to do away with basic notions of Generative Grammar established over the years (cf. Halle 1996, in Burzio 1996, Prince 1996, McMahon 2000). This misunderstanding has been also dealt with by other researchers (Burzio 1996, Prince 1996). The virtue of this chapter is that it straightforwardly addresses the differences between OT and other theories of UG, revealing OT's advantages over other models, which made it the most prominent theory of mid 90's till today. This is again something very authentic compared to other OT textbooks, where one cannot really find this kind of direct comparison between OT and other approaches.
Chapter 3. The Results of Optimality Theory
This chapter might be considered to constitute the main bulk the book, since the formal consequences directly related to OT are presented here. The author goes back to the premises of OT, which were illustrated and explained in chapter 1, but this time he includes many examples and theoretical analyses. As a result, this chapter is again divided into subchapters; the first one deals with the consequences of Faithfulness/Markedness Interactions, the second subchapter examines the consequences of constraint violability, and finally the third subchapter explores the consequences of Globality and Parallelism.
In more detail, the first part goes through the details regarding basic concepts and relations, such as the one between 'OT inventories' and 'Richness of the Base' (hereafter ROTB), or the one between 'Lexicon', 'ROTB' and 'Lexicon Optimization'. It is explained why thanks to ROTB and OT the 'duplication problem' does not constitute a problem anymore for phonological theory. The above ideas and argumentation are supported by additional evidence from Syntax. Other notions that are investigated in relation to the faithfulness and markedness interactions are the 'distributional restrictions' of linguistic items (of constraints in the case of OT), the notions of 'factorial typology', 'contextual neutralization', 'conspiracies', and 'harmonic ascent'.
In the second subchapter the combination of OT constraints is discussed with the exemplification of cases influencing both separate modules of grammar and the interface of grammatical components. This part once again gives to the author the chance to compare OT with the P&P and SPE models. The flexibility that OT demonstrates is one of the things that characterize the first but not the latter approaches. I wouldn't characterize his examples anything else, but convincing. McCarthy explains all the cases he presents in depth and a very analytical way.
Finally, in the third subchapter the consequences of Globality and Parallelism are explored by means of examples from both phonology and Syntax. The author goes into challenging problems of OT such as Opacity. He presents the analyses of opacity within OT, namely harmonic serialism (hereafter HS), the PARSE/FILL implementation of faithfulness in OT, correspondence theory, and sympathy. He argues that HS and the other approaches to opacity better account for opacity compared to SPE and Lexical Phonology (hereafter LP). I must say that I am not absolutely convinced as far as this view is concerned. On the one hand, it is obvious that violable constraints better explain opacity compared to rules. Still, HS has clearly inherited the property of derivation from LP. This property definitely contradicts with the OT premises of Parallelism and Globality, though. Consequently, it would be important to see if OT could deal with opacity with its own tools. But since OT needs kind of derivational steps to deal with opacity, I do not understand the reason why OT is a better model than LP. This is certainly not a problem of this book in particular, but OT in general. It is true what McCarthy says, that a lot of further research is needed, in order to solve the opacity problems. We are certainly looking forward to the results of this ongoing research.
Chapter 4. The Connections of Optimality Theory
As we saw, chapter 3 explores phonological, morphological and syntactic issues. Chapter 4 puts forward the 'relations' of OT with other areas of study, and it is not strictly related to theory. It is rather general and aims at providing the reader with basic issues concerning the connection of OT with these other areas. The chapter is roughly divided in 5 parts. In the first part McCarthy tries to find out how the core principles of OT apply in Syntax, since there is not as much research going on in Syntax compared to Phonology and researchers have literally not made up their minds regarding how the core notions of OT work in Syntax.
In the second part, Learnability and acquisition issues are dealt with. The author underlines the fact that OT was immediately related with the study of Learnability issues, something that on the one hand, immediately differentiated OT form other generative frameworks, and on the other hand, gave rise to the learning algorithms that have been proposed by various researchers (among others Tesar and Smolensky 1993, 1998, 2000, Turkel 1994, Boersma 1998). The pioneering work on learning algorithms is another advantage of OT over other frameworks. And I totally agree. The developments of OT in issues of language acquisition are far beyond those suggested by any other theoretical model.
What I find a weak point in this part of this chapter though, is that McCarthy gives emphasis mostly on the R(ecursive) C(onstraint) D(emotion) algorithm. The author highlights the fact that he refers to the RCD even though he is very well aware of the alternative algorithms proposed. Since he realizes that the RCD fails to account for cases of variation in language development, he should have provided information about the algorithms that manage to 'solve' the problem (cf. Boersma 1998, Boersma and Hayes 2001). By not doing that he leaves space for one to assume that OT does not have a solution to the problem. The G(radual) L(earning) A(lgorithm), for example, deals very well with it. Referring to these works only in the suggested readings is kind of unfortunate, I think. For the rest, the remainder of this part nicely relates the predictions of OT with language acquisition. In the short third part, the author demonstrates in a simple but comprehensive way the relation of OT to computation, giving the example of Tesar's work.
The fourth part refers to another accomplishment of OT, namely the bridging of the gap between functionalism and formal grammars. It is convincingly shown how violable OT constraints combine the above competing approaches and can account for 'functional insights about ease of articulation, clarity of perception, and the tension between them into a formal grammar (p. 227)'. Finally, the last part is dedicated to language variation and language change from both a synchronic and diachronic point of view. I would have preferred to see this part after the part on variation in language acquisition. Since the author gives evidence about how scientists have historically accounted for language change and variation and he ends up with the algorithm that Cho (1998) and Antilla and Cho (1998) have proposed, the references on variation in language acquisition and adult languages would nicely constitute a whole. The way it is now, it looks a little bit discontinuous. Nevertheless and all in all, the chapter is very informative and the issues are nicely presented.
In the Epilogue the general goals of the book are summarized and the issues remaining for further research are referred. McCarthy clearly indicates the fact that OT has not found the answer to everything, but research is ongoing.
One of the virtues of the TGOT is that it is a very well structured and well-organized book. The chapters are divided further in subchapters and its ideas are clearly pointed out. Consequently, it's difficult for the reader to be left with questions, and even if he is, appendix A (FAQ) clarifies many things in an eligible and elegant way. Another characteristic is that all chapters can be read individually, still all of them manage to make a unified whole in the end.
Still another unique accomplishment of this book is that it brings together the past and the present of Phonological theory, even if the ultimate goal is to give the credit to OT. As a result, in most of the phenomena examined the author brings up the contributions of the old and new theories and he openly challenges the comparison between them. This helps the reader a lot, I assume, to understand the strong points of past theories that are not overcome and are still valid in new approaches and the weak points that have been considered by new accounts in order for the latter ones to be fully fledged. The author convincingly shows not only "how to do OT" but also "why do OT".
Furthermore, McCarthy covers as many phenomena as possible from all modules of Generative Grammar. Of course, references on phonological phenomena are more extensive, since OT has started out as a model of Phonology. Research in other fields, such as semantics, has started relatively recently. Finally, the book has an extensive, if not exhaustive, bibliography on probably all the issues raised in OT up to now.
I wish to thank J.J. McCarthy for providing me with a list of known errata. You can find this list in the following webpage:
Antilla, A. and Y-M. Y. Cho. 1998. Variation and Change in Optimality Theory. Lingua 104. 31-56.
Boersma, P. 1998. Functional Phonology: Formalizing the Interaction between Articulatory and Perceptual Drives. The Hague: Holland Academic Graphics.
Boersma, P. and B. Hayes. 2001. Empirical Tests of the Gradual Learning Algorithm. Linguistic Inquiry 32. 45-86.
Burzio, L. 1996/2000. The Rise of Optimality Theory. In Cheng, L. and R. Sybesma (eds.). The First Glot International State-of-the-Article Book. The Latest in Linguistics. Mouton de Gruyter. 199-220.
Burzio, L. 1996. A Reply to Prof. Halle. Glot International 2.6.
Cho, Y-M. Y. 1998. Language Change as Reranking of Constraints. In Hogg, R. M. and L. van Bergen (eds.). Historical Linguistics 1995. vol.2. Amsterdam and Philadelphia: John Benjamins. 45-62.
Chomsky, N. and M. Halle. 1968. The Sound Pattern of English. New York: Harper and Row.
Goldsmith, J. 1976. Autosegmental Phonology. Doctoral Dissertation. Cambridge: Massachusetts Institute of Technology.
Hayes, B. 1980. A Metrical Theory of Stress Rules. Doctoral Dissertation. Cambridge, MA: Massachusetts Institute of Technology.
McMahon, A. 2000. Change, Chance and Optimality. Cambrdige: Cambridge University Press.
Prince, A. 1996. A Letter from Alan Prince. Glot International 2.6.
Smolensky, P. 1984. Harmony Theory: Thermal Parallel Models in a Computational Context. In Smolensky, P. and M. S. Riley (eds.). Harmony Theory: Problem Solving, Parallel Cognitive Models, and Thermal Physics. La Jolla: Institute for Cognitive Science, University of California at San Diego.
Stampe, D. 1973. A Dissertation on Natural Processes. Doctoral Dissertation. Chicago: University of Chicago.
Tesar, B. and P. Smolensky. 1993. The Learnability of Optimality Theory. In Aranovich, R., W. Byrne, S. Preuss and M. Senturia (eds.). Proceedings of the Thirteenth West Coast Conference on Formal Linguistics. Stanford, CA: CSLI. 122-137.
Tesar, B. and P. Smolensky. 1998. Learnability in Optimality Theory. Linguistics Inquiry 29. 229-268.
Tesar, B. and P. Smolesky. 2000. Learnability in Optimality Theory. Cambridge, MA: MIT Press.
Turkel, W. 1994. The Acquisition of Optimality Theoretic Systems. Unpublished Manuscript. Vancouver: University of British Columbia.