Matthew W. Crocker, Martin Pickering, and Charles Clifton, Jr.,
eds. (2000) Architectures and Mechanisms for Language
Processing. Cambridge University Press: Cambridge, UK; pp. 361,
GPB 45.00 (ca. $ 67.50 ).
Dieter Hillert, University of California at San Diego
The topic of the volume refers to a discipline that has been
developed in the early 70s when linguistic theories rather than
linguistic observations and descriptions are regarded to be
essential for the examination of cognitive processes. A new
interdisciplinary research field was born, known as
psycholinguistics that explores in linguistic terms the cognitive
dispositions and foundations of language processing (Fodor, 1983;
Frazier, 1979; Foster, 1979). This account should not be confused
with a closely related research field, the "psychology of language",
that applies more traditional concepts used generally in cognitive
psychology (e.g., Bever, 1970). However, since the emergence of
psycholinguistic research, multiple subdisciplines were developed
that cover specific research domains (e.g., computational
psycholinguistics, developmental psycholinguistics,
neurolinguistics).
The volume in review includes mostly topics (but not all) that
refer to the domain of computational psycholinguistics. The editors
consider their book as being a high-level introduction that consists
of a collection of fifteen selected papers named after the first
conference on Architecture and Mechanisms for Language
Processing in 1995 (AMLaP-95, Edinburgh). This European conference
series has been initiated to meet the well-known annual CUNY
Conference on Human Sentence Processing in North-America.
Thus, everybody interested in research on modeling the human
sentence processor may regard this volume as necessary
ingredient of his bookshelf. The contributions focus on the
discussing of different types of models and theories that are
regarded to be highly relevant for describing and/or explaining and
predicting sentence comprehension. These models and theories
represent linguistic, computational, or cognitive accounts and
consider to a certain extent the data of empirical studies conducted
with adult readers. Let us first briefly refer to the work presented in
the volume, before being more evaluative about the research
published.
With exception of Martin Pickering, Charles Clifton, Jr. and
Matthew Crocker's overview chapter ("Architectures and
mechanisms for language processing"; you can read this chapter
as an sample text by following this link
http://uk.cambridge.org/order/WebBook.asp?ISBN=0521631211),
the volume has been divided into four parts each of which consists
of three or four chapters: (1) Frameworks, (2) Syntactic and Lexical
Mechanisms, (3) Syntax and Semantics, and (4) Interpretation.
"Frameworks" can be regarded as a general discussion forum in
which the non-expert has the chance to become more familiar with
the relevant topics. Clifton, Jr.'s paper ("Evaluation models of
human sentence processing") provides a comprehensive overview
of different parsing accounts by referring to various findings
published in the literature. In particular, the reader will be
introduced with the controversial discussion of how people resolve
structural (syntactic) ambiguities by considering the controversial
concept of "modularity". Richard Lewis' contribution ("Specifying
architectures for language processing: Process, control, and
memory in parsing and interpretation") discusses in computational
terms ambiguous sentence structures by considering memory and
control functions. This approach relies in particular on a general
cognitive model developed by Newell (1990: Soar). Michael
Tanenhaus, Michael Spivey-Knowlton and Joy Hanna ("Modeling
thematic and discourse context effects with a multiple constraints
approach: Implications for the architecture of the language
comprehension system") using simulation data to account for
context effects in resolving structural ambiguities. They conclude
that probabilities are intrinsic to the representations themselves.
Finally, Gerry Altman chapter ("Late closure in context: Some
consequences for parsimony) re-evaluates his Referential Theory
that tries to describe the link between sentence structure and
mental world. The second part, "Syntactic and lexical mechanism",
starts with a contribution by Steffan Corley and Matthew Crocker
("Modular statistical hypothesis: Exploring lexical category
ambiguity") who concentrate on statistical probabilities involved in
lexical category disambiguation. In the second chapter of this part
("Lexical syntax and parsing architecture") Merlo and Stevenson
discuss the Competitive Attachment Model that implements
Chomsky's (1981) Government Binding Theory as network
representation. Again, James Henderson outlines in his paper
("Constituency, context, and connectionism in syntactic parsing)
how linguistic (symbolic) structures such as unbound dependencies
might be parsed within a connectionistic framework. The third part,
"Syntax and Semantics", is introduced by Colin Brown and Peter
Hagoort's ("On the electrophysiology of language comprehension:
Implications of the human language system) paper that discusses
the N400 effect (conceptual and/or semantic mismatch") as well as
the P600 effect (syntactic positive shift, SPS). Again, Martin
Pickering and Matthew Traxler ("Parsing and incremental
understanding during reading") presenting their view how parsing
and plausibility interact by considering "testability" as a crucial
factor. Barbara Hemforth, Lars Konieczny and Christoph
Scheepers approach ("Syntactic attachment and anaphor
resolution: The two sides of relative clause attachment") discuss
relative clause attachment and anaphoric binding in light of
German data. The concluding paper by Marcia de Vincenzi
("Cross-linguistic psycholinguistics") mainly discusses from a cross-
linguistic perspective parsing of wh-questions. The final part of this
volume, "Interpretation", opens with a contribution by Lyn Frazier
("On interpretation: Minimal 'minimal lowering'); she adopts
Diesing's (1992) Mapping Hypothesis to account for the resolution
of ambiguous determiner phrases. The chapter next to the last by
Linda Moxey and Anthony Sandford ("Focus effects associated
with negative quantifiers") outlines components of a discourse
model that considers reference assignment of negative quantifiers.
The concluding chapter by Amit Almor ("Theories of anaphor
processing") discusses his Information Load Hypothesis by
considering evidence found with Alzheimer's Disease patients.
It is unfortunate that the publication process took so long.
However, those who like to get a first glance of linguistic research
may require to consult first a textbook before reading this volume.
Otherwise, I enthusiastically recommend this book without any
hesitation to those who like to be become more familiar with
linguistic and computational aspects of parsing. However, insiders
recognize that most authors already published their work in detail
somewhere else; that is, they might be disappointed not to
encounter new ideas and approaches as one might expect from a
new volume. Moreover, the four parts in which the volume has
been subdivided overlap in content too much that these chapters
reflect a comprehensive overview of present-day research within
these domains. Subjectively, I would have preferred a volume in
which parsing strategies are discussed more across different
subdiciplines that consider neuronal networks or evidence from
different kind of populations (an exception represents Amit Almor's
research). Models that are developed at the cutting-edge of a
research paradigm seem to be more promising for innovative findings
with respect to the human sentence processor; and a computational
approach reflects only one tool for exposing the intrinsic cognitive
and neuronal mechanisms that enable humans to parse strings of
words according to rules, preferences, and probabilities. A crucial
part of psycholinguistic research discusses the relationship
between empirical methodology, data, and interpretation. The
parsing evidence presented here relies mostly on findings how
subjects analyze written material. However, linguistic processing
works highly automated and efficient in milliseconds. Thus,
explaining the time window between spoken and written processes
may be crucial for modeling parsing strategies. (Phonological processes
per se may be much faster than reading processes.) Moreover, modern
parsing theories must account for multi-typological phenomena (an
exception represents Marcia de Vincenzi's contribution; e.g, Hillert,
1998). Without devalue each single contribution, personally I was
particularly interested in the approaches by Richard Lewis and by
Paola Merlo and Suzanne Stevenson: Lewis works with different
kind of theories and accounts (e.g., X-bar, Soar) that are linguistic,
computational, and cognitive in nature to account for parsing data;
Merlo and Stevenson try to incorporate X-bar templates into
network representations to account for attachment processes.
Both approaches can be regarded to be innovative. Overall, I truly
consider, as mentioned before, this volume as an important
contribution to the investigation of human sentence processing.
References
Bever, T.G. (1970). The cognitive basis for linguistic structures. In
J.R. Hayes (ed), Cognition and the development of language.
New York: Wiley.
Chomsky, N. (1981). Lectures on government and binding.
Dordrecht: Foris.
Diesing, M. (1992). Indefinites. Cambridge, MA: MIT Press.
Fodor, J.A. (1983). The modularity of mind. Cambridge, MA: MIT
Press.
Foster, K. (1979). Levels of processing and the structure of the
language processor. In W.E. Cooper & E. Walker (eds.),
Sentence processing: Psycholinguistic studies presented to
Merrill Garrett. Hillsdale, NJ: Erlbaum.
Frazier, L. (1979). On comprehending sentences: Syntactic parsing
strategies. West Bend, IN: Indiana University Linguistic Club.
Hillert, D. (1998 ed.), Syntax and Semantics, Vol. 31: Sentence
Processing: A Crosslinguistic Perspective. San Diego:
Academic Press.
Newell, A. (1990). Unified theories of cognition. Harvard University
Press.
[No biography of reviewer provided--ed.]
|
