Unofficial Links and Notes on LFG/OT

Lexical-Functional Grammar. An Introduction to Parallel Constraint-Based Syntax, a textbook by Yehuda Falk  to be published by CSLI, is now available on the web for comments: http://pluto.mscc.huji.ac.il/~msyfalk/Textbook.html.  It looks to me like a very well done introduction, with a focus on English syntax and a comparative perspective toward other generative theories.   I'm impressed!

OT Semantics links.  Check out the website of the Conference on the Optimization of Interpretation at the University of Utrecht (January 2000) for names of some of the people now working in this area.  Brady Clark (p.c.) has tracked down some of the interesting recent work in this area downloadable from the web:

Gerhard Jaeger [with Reinhard Blutner]. 1999. Against lexical decomposition in syntax. to appear in Proceedings of IATL 15.

Gerhard Jaeger. 2000. Optimal syntax and optimal semantics.  Slides from a talk at the University of California at Los Angeles.

Reinhard Blutner [with G. Jäger].  1999.  Competition and interpretation: The German adverbs of repetition .

Reinhard Blutner. 2000. Some Aspects of Optimality in Natural Language Interpretation.  First draft [Optimality Workshop  Utrecht January 2000]

Paul Dekker and Robert van Rooy.  Optimality Theory and GameTheory: Some Parallels.

Helen de Hoop and Henriette de Swart. 1998.  Temporal adjunct clauses in Optimality Theory. MS. OTS Utrecht.

Helen de Hoop. 2000 Optional Scrambling and Interpretation, in H. Bennis, M. Everaert, E. Reuland (eds.),  Interface Strategies, KNAW, Amsterdam, 153-168.

Petra Hendriks and Helen de Hoop. To appear. Optimal theoretic semantics.  In Linguistics and Philosophy.

D. Terence Langendoen. 2000.   An Optimality Theoretic account of the scope of operators.  MS. University of Arizona.

Benjamin Lyngfelt.  2000. OT semantics and control. MS.  University of Gothenburg.

The LFG2000 program is now on the web!

Check out the new webpage by Hanjung Lee which has her two very nice recent OT papers in downloadable form:

 Hanjung Lee.  2000b.  Markedness and Word Order Freezing.  To appear in Peter Sells (ed.), Formal and Empirical Issues in Optimality-theoretic Syntax. Stanford: CSLI Publications.  (Expanded and revised version of the 1999d paper)

Hanjung Lee.  2000c.  Markedness and Pronoun Incorporation. 2000c. To appear in BLS 26.

Andy Way, 1999. A Hybrid Architecture for Robust MT using LFG-DOP. Journal of Experimental and Theoretical Artificial Intelligence 11 (Special Issue on Memory-Based Language Processing)

Rens Bod and Ronald Kaplan, 1998. A Probabilistic Corpus-Driven Model for Lexical-Functional Analysis. Proceedings of ACL/COLING '98, Montreal, Canada.

I've read a fascinating new paper by Nick Evans on agreement and pronominal incorporation:, which argues that what is needed for many polysynthetic languages are "formalisms capable of representing the appropriate level of semantic specificity contributed by the pronominal affix, then (where necessary) unifying this with relevant information contributed by external nominals."

Nicholas Evans. 1999. Why argument affixes in polysynthetic languages are not pronouns: evidence from Bininj Gun-wok. Sprachtypol. Univ. Forsch. (STUF), Berlin 52, 255-281.

Nigel Vincent. 1999. The evolution of c-structure: prepositions and PPs from Indo-European to Romance. Linguistics 37-6, 1111-1153.

I sat down to update the OT-LFG page, which had been sadly neglected since October, and I was amazed to see how much work there has been in this area.    Check out the links to new work by Aaron Broadwell, Hye-Won Choi, Jonas Kuhn, Hanjung Lee, Helge Lødrup , Yukiko Morimoto, Peter Sells, Devyani Sharma, and Nigel Vincent--not to mention the papers by  Anette Frank et al., Miriam Butt, Joan Bresnan, and Ash Asudeh. I've probably left out other important (downloadable) OT-LFG papers.  Please don't hesistate to contact me to let me know, so I can catch up.

In connection with OT, I must mention how much I've learned from reading and thinking about the new paper on the Gradual Learning Algorithm by Paul Boersma and Bruce Hayes:

Paul Boersma and Bruce Hayes. 1999. Empirical tests of the Gradual Learning Algorithm. ROA
http://ruccs.rutgers.edu/roa.html and Paul Boersma's  and Bruce Hayes' web pages.

Those interested in morphology and LFG will want to read the new paper by Andrew Spencer and Louisa Sadler, which can be downloaded from Louisa's site and from Andy's personal homepage as well.  This paper develops a formally explicit realizational theory of morphology for LFG, arguing for the separation of f-structure and m-structure features in the analysis of the Latin periiphrasis construction discussed earlier by Kersti Börjars and Nigel Vincent:

Andrew Spencer and Louisa Sadler. 1999. Syntax as an exponent of morphological features.   DRAFT of December 1999 (under editorial review).

Andrew Spencer.  n.d. Analytic tenses in Slavic - an LFG approach.

B. Cormons, 1999. Analyse et desambiguisation: Une approche a base de corpus (Data-Oriented Parsing) pour les representations lexicales fonctionnelles. PhD thesis, Universite de Rennes, France.

A. Way, 1999. "A Hybrid Archtecture for Robust MT using LFG-DOP", Journal of Experimental and Theoretical Artificial Intelligence 11 (Special Issueon Memory-Based Language Processing).

Links up to August 1999 are now on a separate page.
Links up to August 1998 are on a separate page.

Yes, there are some official links, too: start with Essex LFG and Stanford LFG, to find more internet-accessible information about current research, publications, people, and events.

Helge Lødrup 's beautiful paper given at LFG98 has now been published as--

 Helge Lødrup. 1999.  ``Inalienables in Norwegian and Binding Theory.'' Linguistics 37, 3, Pp. 365 - 388.

``Linking and Optimality in the Norwegian Presentational Focus Construction.'' Forthcoming

``Exceptions to the Norwegian passive: Unaccusativity, aspect and thematic roles.'' Forthcoming

Judith Berman's paper  ``Does German Satisfy the Subject Condition? '' is now published in the LFG99 on-line proceedings.  This is one of the fascinating papers which I have recommended to my students.  Here's the abstract:

In this paper I argue that German satisfies the Subject Condition even if it has clauses without a (c-structure) subject-constituent. In particular, I argue that expletive subjects are provided by the verbal agreement morphology; from this it follows that the insertion of an expletive subject in c-structure is not necessary to satisfy the Subject Condition and therfore by Economy of Expression is prohibited.

Check out the new Optimal Typology Project webpage.  This page includes downloadable papers and information about seminars, workshops, and research activities relating to OT syntax at Stanford and UCSC.

Mark Johnson's website is always worth checking out for new research results of interest .    The latest addition is a  new paper on defining a probability distribution for LFGs, which should enable new statistical NLP work incorporating LFGs:

M. Johnson, S. Geman, S. Canon, Z. Chi and S. Riezler (1999) ``Estimators for Stochastic ``Unification-based'' Grammars'' to appear in The Proceedings of the ACL 1999.

Peter Sells has updated his website with more downloadable LFG-related papers, including some of his new work on  Scandinavian word order in OT-LFG.   Peter's work is attracting a lot of international attention these days (he has been tapped to give several keynote addresses in 2000), but not everyone knows that he leads a double life: he is not only Peter Sells the linguist, but Pells, the international bass guitarist -

Tetsuro Nishino,  a theoretical computer scientist at The University of Electro-Communications  in Tokyo, does research on quantum computation, neuroidal nets and circuit complexity theory.   He has recently put up an English language website at  http://www.sw.cas.uec.ac.jp/tnlab/member/nishino.html, which contains references to his work on the mathematical analysis of Lexical-Functional Grammars and his new work on Optimality Theoretic LFGs on Neuroidal Nets.

Devyani Sharma's LFG99 paper on constructive morphology in Hindi is now available in html  and postscripton her website.  It will appear with the LFG99 proceedings , which are not on-line yet.

Many have been waiting to get a copy of Judith Aissen's stimulating and fruitful paper  Markedness and Subject Choice in Optimality Theory.    A draft of the paper is finally available in an rtf version at her newly developing website at http://ling.ucsc.edu/~aissen/.

Don't miss the following three exciting new OT-LFG papers.  For more details and links, see the OT-LFG page, which I maintain.

        Ash Asudeh. 1999. Linking, optionality, and ambiguity in Marathi: An Optimality Theory analysis.  Ms. Stanford University Department of Linguistics.  Postscript and pdf.

        Jonas Kuhn. 1999. Two ways of formalizing OT Syntax in the LFG framework.  Ms. Institut für maschinelle Sprachverarbeitung, Universität Stuttgart.

        Hanjung Lee. 1999. Markedness and Word Order Freezing: An OT Account. Ms. Stanford University Department of Linguistics.
 

Thanks to the excellent efforts of Professor Michiko Nakano, AILA 99 - The 12th World Congress of Applied Linguistics on August 1 - 6, 1999 had a feast of LFG : two keynotes and four symposia were given on LFG and related topics.    The keynote addresses were by myself and Ron Kaplan, on theoretical and computational linguistics, respectively. Of the four symposia, two offered introductions to LFG and to recent research in LFG.  Peter Sells, Yo Matsumoto, Tara Mohanan and Ron Kaplan were contributors to these, each presenting an excellent paper motivating the architecture of LFG.   In the advanced symposium, Peter presented his insightful work on Japanese Postposing using constructive morphology, and Ron presented his joint work with Rens Bod on DOP-LFG, an  exemplar-based rather than generalization-based theory of grammar. Then there was a  very well-attended and well received Symposium on Lexicalist Syntactic Theories and Language Education, starting with an  excellent overview by Peter Sells, going on to Ron Kaplan's discussion of an LFG-based computer-aided language learning program, and to Michiko Nakano's review of experimental results on dative constructions and negative auxiliaries.   Finally came the Symposium on LFG and Its Computational Implications, with a most useful mathematical overview by Professor Etsuro Moriya (featuring the equivalence of a polynomial-recognizable subclass of LFGs with linear index grammars, head grammars, tree adjoining grammars, combinatory categorial grammars, and linear CF tree grammars) and a remarkable paper on Optimality Theoretic LFG on Neuroidal Nets by Professor Tetsuro Nishino using a dynamic, parallel  computational model to evaluate the OT-LFG candidate space. There are plans to publish the collection of papers resulting from this conference.  Copies of the papers will soon be available.  For more information contact Professor Nakano or the individual presenters.

Rens Bod's book, Beyond Grammar.  An Experience-Based Theory of Language, has now been published by CSLI Publications .    This book is of interest not only for its  technical proposals for integrating formal language theory with statistical linguistics, but also for its exemplar-based epistemology.  Knowledge of language, according to this theory, should be understood not as a grammar, but as a "statistical ensemble of language experiences that changes slightly every time a new utterance is perceived or produced".  The last chapter (written with Ron Kaplan) is about DOP-LFG.   You can find out more about DOP -LFG  on the Stanford LFG web site: http://www-lfg.stanford.edu/lfg/.

Here are some brief notes on themes and highlights from LFG99.

One highlight of LFG99  took place on Day One, with Helge Dyvik's incisive , lucid, and graceful paper on the universality of f-structures and the use of translation equivalents as a domain for semantics. In the midst of the "applied" Workshop on Grammar Development  (beautifully organized by Victoria Rosén) Helge raised one of the most profound  foundational questions of the conference.  Annie Zaenen's talk  in the same workshop on the social and intellectual problems of managing a  multilingual grammar development project was also very well received. The workshop included a  demonstration of very high-quality  automatic translation by transfer  from English to French by  Anette Frank and a debate among members of the PARGRAMproject  on the analysis of auxiliaries and the criteria for determining f-structures.

Day Two provided another highlight: the superb Workshop on Native American Languages organized by Aaron Broadwell.  We were treated to  OT-, OT-LFG-, and LFG-oriented analyses of four languages by researchers who had obtained  all of the primary data  by their own field work: Jack Martin on Creek, Cathy O'Connor on Northern Pomo, Aaron Broadwell on two Otomanguean languages, and Amy Dahlstrom  on Fox. The papers were all packed with fascinating data, analyses, and insights.  Those who work on "technogrammar" (by which I mean grammar formalisms and architectures) may not appreciate how rare, difficult, and valuable it is to have research of the kind represented by this workshop.  Aaron Broadwell's OT-LFG analysis of pied piping in terms of edge-alignment was an eye-opening alternative to the conventional movement approaches. Cathy O'Connor's discussion of kinship nouns in Pomo  using harmonic alignment of syntactic markedness hierarchies was her first foray into OT-LFG.  Speaking with disarming directness and humor, Cathy presented an insightful and thought-provoking analysis that linked a small corner of Pomo grammar to pervasive typological generalizations. One participant told me that  this workshop left her on a high of intellectual exhilaration.

I thought the two computationally oriented talks on Day Two were particularly impressive.  The first was by Jonas Kuhn on "An Architecture for Structure-Function Mapping", in which he pointed out the advantages and problems posed by using highly theoretical endocentric principles of c-structure including the extended head theory  and economy of expression I have proposed in my book ms. and elsewhere.   His attack on the problem looks extremely promising.  One knowledgeable member of the audience told me that Jonas' paper was the high point of the conference for him.
The second  computational paper, by Josef van Genabith, Andy Way, and Louisa Sadler, was excitingly presented in Josef van Genabith's dynamic  and passionate style, inviting and compeling the audience to follow the crystal clear logic of the argument. The paper outlined how ordinary treebanks - corpora tagged with part of speech labels and  syntactic brackets - could be used to automatically extract f-structures,  induce c-structure rules and lexical forms, and  express generalizations by compacting the grammars in a structure-preserving way.  Closely connected to Andy Way's Essex Ph.D. dissertation topic on DOP-LFG machine translation , this project shows how impressively quickly computational linguistics is advancing.

A notable theme of the LFG99 conference was the presence of typologically oriented talks.  Examples included Peter Austin's paper on  the implications of clause-linking case morphology for constructive morphology, Ackerman and Moore's well-constructed paper on the morphology of telicity,  and Anna Siewierska's investigation of the alignment of pronominal forms with the relational hierarchy which brings her extensive database of pronominal systems into contact with OT-LFG.  A number of  notable typologists were in the audience as well.

I haven't done justice to the other papers presented.  The on-line publications will reflect the conference better than these unofficial notes. At what other conference, I wondered, could you find such a mixture of theoretical, empirical, and computational research? Many commented on the atmosphere of intellectual openness, and the richness of perspectives and ideas represented.

More LFG-related events took place this summer.  See the UIUC LSA Summer Institute and  ESSLLI99 websites.

I hope to put up some information about the results starting to come out of my Optimal Typology project (based on collaborative research with Judith Aissen) before too long....

One additional bit of news that not everyone seems to have gotten:  Chris Manning will  be joining the Departments of Computer Science and Linguistics at Stanford University this fall (1999).  I noted previously that Avery Andrews' and Chris's new book Complex Predicates and Information Spreading in LFG is now in print.  I am finding this fascinating  reading, for the typological variety of complex predicates discussed [including Urdu, Romance, Wagiman (Australia),Tariana, Misumalpa],  for the enlightening overview comparisons of LFG and HPSG analyses of complex predicates, and for the very clearly written exposition of their  own proposed formal framework, which is inspired by ideas from both frameworks.

TTFN