UHH>Informatik>NatS>CDG Web>InternalPages>EssentialReading>SentenceProcessing (17 Oct 2012, UnknownUser) Print version

Sentence Processing

pic-SMR.gif

Keywords: incremental parsing, expectations, psychological evidence, cognitive parsing, optimal parsing, cognitive models for sentence processing, psycholinguistics, garden path

See also: OptimalityTheory, IncrementalComputation, CognitiveNeuroscience, SentenceProcessingReading

Sources:

[BaderEtal04]
Markus Bader, Jana Häussler, and Josef Bayer. Toward an integrated model of structure and frequency. In Architectures and Mechanisms for Language Processing, AMLaP. Laboratoire Parole et Langage, CNRS, Université de Provence, September 2004.
[BudiuAnderson04]
Raluca Budiu and John R. Anderson. Interpretation-based processing: a unified theory of semantic sentence comprehension. In Cognitive Science [Budiu01], pages 1--44. [ .pdf ]
Abstract: We present interpretation-based processing --- a theory of sentence processing that builds a syntactic and a semantic representation for a sentence and assigns an interpretation to the sentence as soon as possible. That interpretation can further participate in comprehension and in lexical processing and is vital for relating the sentence to the prior discourse. Our theory offers a unified account of the processing of literal sentences, metaphoric sentences, and sentences containing semantic illusions. It also explains how text can prime lexical access. We show that word literality is a matter of degree and that the speed and quality of comprehension depend both on how similar words are to their antecedents in the preceding text and how salient the sentence is with respect to the preceding text. Interpretation-based processing also reconciles superficially contradictory findings about the difference in processing times for metaphors and literals. The theory has been implemented in ACT-R (Anderson & Lebiere, 1998)
[KempenHarbusch03]
Gerard Kempen and Karin Harbusch. An artificial opposition between grammaticality and frequency: comment on [bornkessel, schlesewsky, and friederici (2002)]. In Cognition [BornkesselEtal02], pages 205--210. [ .pdf ]
Abstract: In a recent Cognition paper (Cognition 85 (2002) B21), Bornkessel, Schlesewsky, and Friederici report ERP data that they claim "show that online processing difficulties induced by word order variations in German cannot be attributed to the relative infrequency of the constructions in question, but rather appear to reflect the application of grammatical principles during parsing" (p. B21). In this commentary we demonstrate that the posited contrast between grammatical principles and construction (in)frequency as sources of parsing problems is artificial because it is based on factually incorrect assumptions about the grammar of German and on inaccurate corpus frequency data concerning the German constructions involved.
[Sturt03]
Patrick Sturt. The time-course of the application of binding constraints in reference resolution. Journal of Memory and Language, 48(3):542--562, 2003. [ .pdf ]
Abstract: We report two experiments which examined the role of binding theory in on-line sentence processing. Participants' eye movements were recorded while they read short texts which included anaphoric references with reflexive anaphors (himself or herself). In each of the experiments, two characters were introduced into the discourse before the anaphor, and only one of these characters was a grammatical antecedent for the anaphor in terms of binding theory. Both experiments showed that Principle A of the binding theory operates at the very earliest stages of processing; early eyemovement measures showed evidence of processing diffculty when the gender of the reflexive anaphor mismatched the stereotypical gender of the grammatical antecedent. However, the gender of the ungrammatical antecedent had no effect on early processing, although it affected processing during later stages in Experiment 1. An additional experiment showed that the gender of the ungrammatical antecedent also affected the likelihood of participants settling on an ungrammatical final interpretation. The results are interpreted in relation to the notions of bonding and resolution in reference processing.
[McKoonRatcliff03]
Gail McKoon and Roger Ratcliff. Meaning through syntax: Language comprehension and the reduced relative clause construction. Psychological Review, 110(3):490--525, 2003. [ .pdf ]
Abstract: A new explanation is proposed for a long standing question in psycholinguistics: Why are some reduced relative clauses so difficult to comprehend? It is proposed that the meanings of some verbs like race are incompatible with the meaning of the reduced relative clause and that this incompatibility makes sentences like The horse raced past the barn fell unacceptable. In support of their hypotheses, the authors show that reduced relatives of The horse raced past the barn fell type occur in naturally produced sentences with a near-zero probability, whereas reduced relatives with other verbs occur with a probability of about 1 in 20. The authors also support the hypotheses with a number of psycholinguistic experiments and corpus studies.
[KamideEtal03]
Yuki Kamide, Gerry Altmann, and Sarah L. Haywood. The time-course of prediction in incremental sentence processing: Evidence from anticipatory eye movements. Journal of Memory and Language, 49:133--156, 2003. see Corrigendum. [ .pdf ]
Abstract: Three eye-tracking experiments using the visual-world paradigm are described that explore the basis by which thematic dependencies can be evaluated in advance of linguistic input that unambiguously signals those dependencies. Following Altmann and Kamide (1999), who found that selectional information conveyed by a verb can be used to anticipate an upcoming Theme, we attempt to draw here a more precise picture of the basis for such anticipatory processing. Our data from two studies in English and one in Japanese suggest that (a) verb-based information is not limited to anticipating the immediately following (grammatical) object, but can also anticipate later occurring objects (e.g., Goals), (b) in combination with information conveyed by the verb, a pre-verbal argument (Agent) can constrain the anticipation of a subsequent Theme, and (c) in a head-final construction such as that typically found in Japanese, both syntactic and semantic constraints extracted from pre-verbal arguments can enable the anticipation, in effect, of a further forthcoming argument in the absence of their head (the verb). We suggest that such processing is the hallmark of an incremental processor that is able to draw on different sources of information (some non-linguistic) at the earliest possible opportunity to establish the fullest possible interpretation of the input at each moment in time.
[HopfEtal03]
Jens-Max Hopf, Markus Bader, Michael Meng, and Josef Bayer. Is human sentence parsing serial or parallel? evidence from event-related brain potentials. Cognitive Brain Research, 15:165--177, 2003. [ .pdf ]
Abstract: In this ERP study we investigate the processes that occur in syntactically ambiguous German sentences at the point of disambiguation. Whereas most psycholinguistic theories agree on the view that processing difficulties arise when parsing preferences are disconfirmed (so-called garden-path effects), important differences exist with respect to theoretical assumptions about the parser s recovery from a misparse. A key distinction can be made between parsers that compute all alternative syntactic structures in parallel (parallel parsers) and parsers that compute only a single preferred analysis (serial parsers). To distinguish empirically between parallel and serial parsing models, we compare ERP responses to garden-path sentences with ERP responses to truly ungrammatical sentences. Garden-path sentences contain a temporary and ultimately curable ungrammaticality, whereas truly ungrammatical sentences remain so permanently a difference which gives rise to different predictions in the two classes of parsing architectures. At the disambiguating word, ERPs in both sentence types show negative shifts of similar onset latency, amplitude, and scalp distribution in an initial time window between 300 and 500 ms. In a following time window (500 700 ms), the negative shift to garden-path sentences disappears at right central parietal sites, while it continues in permanently ungrammatical sentences. These data are taken as evidence for a strictly serial parser. The absence of a difference in the early time window indicates that temporary and permanent ungrammaticalities trigger the same kind of parsing responses. Later differences can be related to successful reanalysis in garden-path but not in ungrammatical sentences.
[Ferreira03]
Fernanda Ferreira. The misinterpretation of noncanonical sentences. Cognitive Psychology, 47(2):164--203, 2003. [ .pdf ]
Abstract: Research on language comprehension has focused on the resolution of syntactic ambiguities, and most studies have employed garden-path sentences to determine the system's preferences and to assess its use of nonsyntactic sources information. A topic that has been neglected is how syntactically challenging but essentially unambiguous sentences are processed, including passives and object-clefts sentences that require thematic roles to be assigned in an atypical order. The three experiments described here tested the idea that sentences are processed both algorithmically and heuristically. Sentences were presented aurally and the participants task was to identify the thematic roles in the sentence (e.g., Who was the do-er?). The rst experiment demonstrates that passives are frequently and systematically misinterpreted, especially when they express implausible ideas. The second shows that the surface frequency of a syntactic form does not determine ease of processing, as active sentences and subject-clefts were comprehended equally easily despite the rareness of the latter type. The third experiment compares the processing of subject- and object-clefts, and the results show that they are similar to actives and passives, respectively, again despite the infrequent occurrence in English of any type of cleft. The results of the three experiments suggest that a comprehensive theory of language comprehension must assume that simple processing heuristics are used during processing in addition to (and perhaps sometimes instead of) syntactic algorithms. Moreover, the experiments support the idea that language processing is often based on shallow processing, yielding a merely good enough rather than a detailed linguistic representation of an utterance s meaning.
[BornkesselEtal02]
Ina Bornkessel, Matthias Schlesewsky, and Angela D. Friederici. Grammar overrides frequency: evidence from the online processing of flexible word order. Cognition, 85(2):B21--B30, September 2002. [ .pdf ]
Abstract: We show that online processing difficulties induced by word order variations in German cannot be attributed to the relative infrequency of the constructions in question, but rather appear to reflect the application of grammatical principles during parsing. Event-related brain potentials revealed that dative-marked objects in the initial position of an embedded sentence do not elicit a neurophysiologically distinct response from subjects, whereas accusative-marked objects do. These differences are predictable on the basis of grammatical distinctions (i.e. underlying linguistic properties), but not on the basis of frequency information (i.e. a superficial linguistic property). We therefore conclude that the former, but not the latter, guides syntactic integration during online parsing.
[Crocker02]
Matthew W. Crocker. Review of ”sentence comprehension: The integration of habits and rules”. In Computational Linguistics [TownsendBever01], pages 238 -- 241.
[TraxlerEtal02]
Matthew J. Traxler, Robin K. Morris, and Rachel E. Seely. Processing subject and object relative clauses: Evidence from eye-movements. Journal of Memory and Language, 47:69--90, 2002. [ .pdf ]
Abstract: Three eye-movement-monitoring experiments investigated processing of sentences containing subject-relative and object-relative clauses. The first experiment showed that sentences containing object-relative clauses were more difficult to process than sentences containing subject-relative clauses during the relative clause and the matrix verb. The second experiment manipulated the plausibility of the sentential subject and the noun within the relative clause as the agent of the action represented by the verb in the relative clause. Readers experienced greater difficulty during processing of sentences containing object-relative clauses than subject-relative clauses. The third experiment manipulated the animacy of the sentential subject and the noun within the relative clause. This experiment demonstrated that the difficulty associated with object-relative clauses was greatly reduced when the sentential subject was inanimate. We interpret the results with respect to theories of syntactic parsing.
[SpiveyEtal02]
Michael J. Spivey, Michael K. Tanenhaus, Kathleen M. Eberhard, and Julie C. Sedivy. Eye movements and spoken language comprehension: Effects of visual context on syntactic ambiguity resolution. Cognitive Psychology, 45(4):447--481, 2002. [ .pdf ]
Abstract: When participants follow spoken instructions to pick up and move objects in a visual workspace, their eye movements to the objects are closely time-locked to referential expressions in the instructions. Two experiments used this methodology to investigate the processing of the temporary ambiguities that arise because spoken language unfolds over time. Experiment 1 examined the processing of sentences with a temporarily ambiguous prepositional phrase (e.g., Put the apple on the towel in the box) using visual contexts that supported either the normally preferred initial interpretation (the apple should be put on the towel) or the less-preferred interpretation (the apple is already on the towel and should be put in the box). Eye movement patterns clearly established that the initial interpretation of the ambiguous phrase was the one consistent with the context. Experiment 2 replicated these results using prerecorded digitized speech to eliminate any possibility of prosodic differences across conditions or experimenter demand. Overall, the ndings are consistent with a broad theoretical framework in which real-time language comprehension immediately takes into account a rich array of relevant nonlinguistic context.
[GrodnerEtal02]
Daniel Grodner, Edward Gibson, and Susanne Tunstall. Syntactic complexity in ambiguity resolution. Journal of Memory and Language, 46(2):267--295, 2002. [ .pdf ]
Abstract: This article presents two self-paced reading experiments which investigate the role of storage costs associated with maintaining incomplete syntactic dependencies in structural ambiguity resolution. We argue that previous work has been equivocal regarding syntactic influences because it has examined ambiguities where there is little or no resource differential between competing alternatives. The candidate structures of the ambiguities explored here incur substantially different storage costs. The results indicate that storage-based biases can be sufficiently powerful to create difficulty for a structural alternative even when it is promoted by nonsyntactic factors. These findings are incorporated into a model of ambiguity resolution in which structural biases operate as independent graded constraints in selecting between structural alternatives.
[FilipEtal02]
Hana Filip, Michael K. Tanenhaus, Gregory N. Carlson, Paul D. Allopenna, and Joshua Blatt. Reduced Relatives Judged Hard Require Constraint-Based Analyses, pages 255--280. Sentence Processing and the lexicon: formal, computational, and experimental perspectives. Benjamins, Amsterdam, 2002.
[TownsendBever01]
David J. Townsend and Thomas G. Bever. Sentence comprehension: The integration of habits and rules. Language, speech, and communication series. MIT Press, Cambridge, MA, May 2001. [ http ]
[KellerAsudeh01]
Frank Keller and Ash Asudeh. Constraints on linguistic coreference: Structural vs. pragmatic factors. In Johanna D. Moore and Keith Stenning, editors, Proceedings of the 23rd. Annual Conference of the Cognitive Science Society, pages 483--488, Mahwah, NJ, 2001. Lawrence Erlbaum Associates. [ .pdf ]
Abstract: Binding theory is the component of grammar that regulates the interpretation of noun phrases. Certain syntactic configurations involving picture noun phrases (PNPs) are problematic for the standard formulation of binding theory, which has prompted competing proposals for revisions of the theory. Some authors have proposed an account based on structural constraints, while others have argued that anaphors in PNPs are exempt from binding theory, but subject to pragmatic restrictions. In this paper, we present an experimental study that aims to resolve this dispute. The results show that structural factors govern the binding possibilities in PNPs, while pragmatic factors play only a limited role. However, the structural factors identified differ from the ones standardly assumed.
[FriedericiEtal01]
A. D. Friederici, A. Mecklinger, K. Spencer, K. Steinhauer, and E. Donchin. Syntactic parsing preferences and their on-line revisions: A spatio-temporal analysis of event-related brain potentials. Cognitive Brain Research, 11:305--323, 2001. [ .pdf ]
[ChristiansonEtal01]
Kiel Christianson, Andrew Hollingworth, John F. Halliwell, and Fernanda Ferreira. Thematic roles assigned along the garden path linger. Cognitive Psychology, 42:368--407, 2001. [ .pdf ]
Abstract: In the literature dealing with the reanalysis of garden path sentences such as While the man hunted the deer ran into the woods, it is generally assumed either that people completely repair their initial incorrect syntactic representations to yield a final interpretation whose syntactic structure is fully consistent with the input string or that the parse fails. In a series of five experiments, we explored the possibility that partial reanalyses take place. Specifically, we examined the conditions under which part of the initial incorrect analysis persists at the same time that part of the correct final analysis is constructed. In Experiments 1a and 1b, we found that both the length of the ambiguous region and the plausibility of the ultimate interpretation affected the likelihood that such sentences would be fully reanalyzed. In Experiment 2, we compared garden path sentences with non-garden path sentences and compared performance on two different types of comprehension questions. In Experiments 3a and 3b, we constructed garden path sentences using a small class of syntactically unique verbs to provide converging evidence against the position that people employ some sort of ”general reasoning” or pragmatic inference when faced with syntactically difficult garden paths. The results from these experiments indicate that reanalysis of such sentences is not always complete, so that comprehenders often derive an interpretation for the full sentence in which part of the initial misanalysis persists. We conclude that the goal of language processing is not always to create an idealized structure, but rather to create a representation that is good enough to satisfy the comprehender that an appropriate interpretation has been obtained.
[CarlsonEtal01]
Katy Carlson, Jr. Charles Clifton, and Lyn Frazier. Prosodic boundaries in adjunct attachment. Journal of Memory and Language, 45(1):58--81, 2001. [ .pdf ]
Abstract: Five studies explored the processing of ambiguous sentences like ”Martin maintained that the CEO lied when the investigation started/at the start of the investigation”. The central question was why particular prosodic boundaries have the effects they do. A written questionnaire provided baseline preferences and suggested that clausal adjuncts (”when the investigation started”) receive more high attachments than nonclausal adjuncts (”at the start of the investigation”). Four auditory studies manipulated the prosodic boundary before the adjunct clause and the prosodic boundary between the matrix clause and its complement. They disconfirm every version of an account where only the local boundary before the adjunct is important, whether the account is based on the acoustic magnitude of the boundary or its phonological type (an intermediate boundary characterized by the presence of a phrase accent vs. an intonational phrase boundary characterized by both a phrase accent and a boundary tone). Instead the results support use of the global prosodic context, especially the relative size of the local boundary and the distant boundary.
[Budiu01]
Raluca Budiu. The Role of Background Knowledge in Sentence Processing. Doctoral dissertation, School of Computer Science, Carnegie Mellon University, 2001. [ .pdf ] In this dissertation I describe a cognitive model of sentence processing. The model operates at the semantic level and can apply to verification or comprehension of metaphoric or literal sentences, isolated or embedded in discourse. It uses an incremental search--and--match mechanism to find a long-term--memory referent (interpretation) for an input sentence. The search is guided by cues such as the last few words read or previous tentative interpretations. The process of comprehension produces a propositional representation for the input sentence and also keeps track of local comprehension failures. The model is implemented in the ACT-R framework and offers a scalable solution to the problem of language comprehension: its performance (in terms of speed and accuracy) is roughly invariant to the number of facts held in the long-term memory. Its predictions match data from psycholinguistic studies with human subjects. Specifically, the sentence-processing model can simulate the comprehension and verification of metaphoric and literal sentences, metaphor-position effects on sentence comprehension, semantic illusions and their dependence on semantic similarity between the distortion and the undistorted term. The products of the sentence-processing model can explain the pattern of sentence recall in text-memory experiments. This dissertation also explores the modeling alternatives faced by the design of a sentence-processing model. I show that, to achieve comprehension speed comparable to that of humans, a model must minimize the explicit search process and rely on semantic associations among words. I also investigate how the representation chosen for propositions and meanings affects the comprehension process in a production-system framework such as ACT-R.
[AlmorEtal01]
Amit Almor, Maryellen C. MacDonald, Daniel Kempler, Elaine S. Andersen, and Lorraine K. Tyler. Comprehension of long distance number agreement in probable alzheimer's disease. Language and Cognitive Processes, 16(1), 2001. [ .pdf ]
Abstract: Two cross-modal naming experiments examined the role of working memory in processing sentences and discourses of various lengths. In Experiment 1, 10 memory impaired patients with probable Alzheimer s disease (AD) and
            1. healthy elderly control participants showed similar sensitivity to violations of subject-verb number agreement in a short sentence condition and similar degradation to this sensitivity in a long sentence condition. Performance in neither length condition correlated with performance on working memory tasks, suggesting that the processes involved in interpreting a grammatical dependency between adjacent and nonadjacent elements are different from those required in the working memory tasks. In Experiment 2, the same 10 AD patients were less sensitive than the 10 control participants to pronounantecedent number agreement violations in a short discourse condition, but neither group was affected by additional length. In this experiment, performance in both the short and long conditions correlated with working memory performance. These results show that grammatical and discourse dependencies pose different memory and processing demands, and that these differences are not simply due to differences in the amount of intervening material between dependent words. The results also suggest that while the working memory de cits characteristic of AD do not interfere with on-line grammatical processing within sentences, they do compromise on-line discourse processing across sentences.
[CrockerBrants00]
Matthew Crocker and Thorsten Brants. Wide-coverage probabilistic sentence processing. Journal of Psycholinguistic Research, 29(6):647--669, November 2000. [ .pdf ]
Abstract: This paper describes a fully implemented, broad coverage model of human syntactic processing. The model uses probabilistic parsing techniques which combine phrase structure, lexical category, and limited subcategory probabilities with an incremental, left-to-right the system to achieve good accuracy on typical, "garden variety" language (i.e. when tested on corpora). Furthermore, the incremental probabilistic ranking of the preferred analyses during parsing also naturally explains observed human behaviour for a range of garden-path structures. We do not make strong psychological claims about the specific probabilistic mechanism discussed here, which is limited by a number of practical considerations. Rather, we argue incremental probabilistic parsing models are, in general, extremely well suited to explaining this dual nature - generally good and occasionally pathological - of human linguistic performance.
[PearlmutterMendelson00]
Neal J. Pearlmutter and A. A. Mendelson. Serial versus parallel sentence comprehension. Manuscript in revision, August 31 2000. [ .ps ]
[SchaferEtal00]
Amy J. Schafer, Shari R. Speer, Paul Warren, and S. David White. Intonational disambiguation in sentence production and comprehension. Journal of Psycholinguistic Research, 29:169--182, 2000. [ .pdf ]
Abstract: Speakers prosodic marking of syntactic constituency is often measured in sentence reading tasks that lack realistic situational constraints on speaking. Results from such studies can be criticized because the pragmatic goals of readers differ dramatically from those of speakers in typical conversation. On the other hand, recordings of unscripted speech do not readily yield the carefully controlled contrasts required for many research purposes. Our research employs a cooperative game task, in which two speakers use utterances from a predetermined set to negotiate moves around gameboards. Results from a set of early versus late closure ambiguities suggest that speakers signal this syntactic difference with prosody even when the utterance context fully disambiguates the structure. Phonetic and phonological analyses show reliable prosodic disambiguation in speakers productions; results of a comprehension task indicate that listeners can successfully use prosodic cues to categorize syntactically ambiguous fragments as portions of early or late closure utterances.
[PickeringEtal00]
Martin J. Pickering, Matthew J. Traxler, and Matthew W. Crocker. Ambiguity resolution in sentence processing: Evidence against frequency-based accounts. Journal of Memory and Language, 43(3):447--475, 2000. [ .pdf ]
Abstract: Three eye-tracking experiments investigated two frequency-based processing accounts: the serial lexical-guidance account, in which people adopt the analysis compatible with the most likely subcategorization of a verb; and the serial-likelihood account, in which people adopt the analysis that they would regard as the most likely analysis, given the information available at the point of ambiguity. The results demonstrate that neither of these accounts explains readers performance. Instead people preferred to attach noun phrases as arguments of verbs even when such analyses were unlikely to be correct. We suggest that these results fit well with a model in which the processor initially favors informative analyses.
[BrantsCrocker00]
Thorsten Brants and Matthew Crocker. Probabilistic parsing and psychological plausibility. In Proceedings of the 18th International Conference on Computational Linguistics, 2000. [ .ps.gz ]
Abstract: Given the recent evidence for probabilistic mechanisms in models of human ambiguity resolution, this paper investigates the plausibility of exploiting current wide-coverage, probabilistic parsing techniques to model human linguistic performance. In particular, we investigate the performance of standard stochastic parsers when they are revised to operate incrementally, and with reduced memory resources. We present techniques for ranking and filtering analyses, together with experimental results. Our results confirm that stochastic parsers which adhere to these psychologically motivated constraints achieve good performance. Memory can be reduced down to 1 exhausitve search) without reducing recall and precision. Additionally, these models exhibit substantially faster performance. Finally, we argue that this general result is likely to hold for more sophisticated, and psycholinguistically plausible, probabilistic parsing models.
[GibsonPearlmutter98]
E. Gibson and Neal J. Pearlmutter. Constraints on sentence comprehension. Trends in Cognitive Sciences, 2:262--268, May 15 1998. [ .ps ]
[CristeaWebber97]
D. Cristea and Bonie Lynn Webber. Expectations in incremental discourse processing. In Proc. 35th Annual Meeting of the Association for Comptutational Linguistics, Madrid, July 1997. [ .html ]
[Phillips96]
Colin Phillips. Order and Structure. PhD thesis, MIT, September 1996. [ .ps ] The aim of this thesis is to argue for the following two main points. First, that grammars of natural language construct sentences in a strictly left-to-right fashion, i.e. starting at the beginning of the sentence and ending at the end. Second, that there is no distinction between the grammar and the parser. In the area of phrase structure, I show that the left-right derivations forced by the principle Merge Right can account for the apparent contradictions that different tests of constituency show, and that they also provide an explanation for why the different tests yield the results that they do. Diagnostics discussed include coordination, movement, ellipsis, binding, right node raising and scope. I present a preliminary account of the interface of phonology and morphology with syntax based on left-right derivations. I show that this approach to morphosyntax allows for a uniform account of locality in head movement and clitic placement, explains certain directional asymmetries in phonology-syntax mismatches and head movement, and allows for a tighter connection between syntactic and phonological phrases than commonly assumed. In parsing I argue that a wide range of structural biases in ambiguity resolution can be accounted for by the single principle Branch Right, which favors building right-branching structures wherever possible. Evidence from novel and existing experimental work is presented which shows that Branch Right has broader empirical coverage than other proposed structural parsing principles. Moreover, Branch Right is not a parsing-specific principle: it is independently motivated as an economy principle of syntax in the chapters on syntax. The combination of these results from syntax and parsing makes it possible to claim that the parser and the grammar are identical. The possibility that the parser and the grammar were identical or extremely similar was explored in the early 1960s, but is widely considered to have been discredited by the end of that decade. I show that arguments against this model which were once valid no longer apply given left-to-right syntax and the view of the parser proposed here.
[MacDonald94]
Maryellen C. MacDonald. Probabilistic constraints and syntactic ambiguity resolution. Language and Cognitive processes, 9:157--201, 1994. Signature R30239. [ .pdf ]
[Lombardo92]
Vincenzo Lombardo. Incremental dependency parsing. In Meeting of the Association for Computational Linguistics, pages 291--293, 1992. [ .pdf ]
[PriceEtal91]
Patti Price, Mari Ostendorf, Stefanie Shattuck-Hufnagel, and C. Fong. The use of prosody in syntactic disambiguation. Journal of the Acoustical Society of America, 90:2956--2970, 1991.
Abstract: Prosodic structure and syntactic structure are not identical; neither are they unrelated. Knowing when and how the two correspond could yield better quality speech synthesis, could aid in the disambiguation of competing syntactic hypotheses in speech understanding, and could lead to a more comprehensive view of human speech processing. In a set of experiments involving 35 pairs of phonetically similar sentences representing seven types of structural contrasts, the perceptual evidence shows that some, but not all, of the pairs can be disambiguated on the basis of prosodic differences. The phonological evidence relates the disambiguation primarily to boundary phenomena, although prominences sometimes play a role. Finally, phonetic analyses describing the attributes of these phonological markers indicate the importance of both absolute and relative measures.

Paul Gorrell (1995): Syntax and parsing.

Cambridge: Cambridge University Press. 1995.

Book Review: (Martin Volk: A review of "Syntax and Parsing" by Paul Gorrell. In: Literary and Linguistic Computing. Vol.11 (2). 1996. 104-105)

The title of this book is perhaps misleading. As a computational linguist I had expected the book to be about computational approaches to natural language parsing. But this book deals with cognitive parsing, i.e. modelling the parsing processes within the human brain. The syntax background is provided by Government Binding (GB) Theory.

Not being an expert in that area I decided to review the book from the perspective of an interested researcher from a neighboring field, trying to learn about GB and its role in human language processing and in cognitive linguistics.

Paul Gorrell outlines a model of human sentence processing from a GB perspective. In this model he distinguishes between primary and secondary operations. Primary operations include building up a phrase structure tree including dominance and precedence relations. These operations are deterministic, i.e. they are not redone because of ambiguities not even in the case of garden path phenomena. In contrast, secondary operations are not deterministic. These include government, c-command, theta assignment, case assignment, and binding. These operations serve to interpret the structure being established by the primary operations. When encountering material that requires reanalysis of interpretation the secondary operations are redone.

Gorrell takes great care to argue that this model is in accordance with empirical findings in cognitive linguistics experiments such as reaction time tests, eye movement tests, shadowing tests etc. Some of these tests he has performed himself, but mostly he tries to accommodate his proposal within the findings of many other researchers in the field.

The book is structured into 6 chapters. Chapter 1 `Introduction' sets the goals:
(p.1) It is a central thesis of this book that recent work within Government Binding (GB) theory (Chomsky 1981, and subsequent work) raises questions about the nature of syntactic knowledge that have long concerned researchers into syntactic processing (parsing).

(p.4) The aim of this book is to give as much attention to the form of the grammar as to the form of the parser.

Chapter 2 gives an introduction to GB theory, distinguishing between derivational (with levels: d-structure, s-structure, logical form) and representational (monostratal) GB. Gorrell assumes a monostratal version for his model. Chapter 3 is devoted to `Analyses of previous work'. It is a detailed survey of the research in human sentence parsing. It looks at the field from many different perspectives providing a host of examples showing that processing difficulties are not restricted to the garden path in
The horse raced past the barn fell.

Let me cite three more examples to reflect the flavor of the discussion:
(p.50) Without her contributions would be inadequate.

(p.53) We gave the man the grant proposal was written by last year a copy of this year's proposal.

(p.55) Though George kept on reading the story really bothered him.

Elaborating on chapter 3, chapter 4 is the heart of the book. Here, the author outlines his model, contrasting it to competing approaches and showing briefly how it applies to languages other than English, namely head-final languages like Japanese and German. Chapter 5 looks at how semantic and pragmatic factors influence structure building in the human brain. The chapter leaves me with the impression that this area is still little explored and that it is much harder to design good experiments. Nevertheless, I agree with Gorrell that issues such as absolute and relative frequency of word occurrence, as well as priming by context are keys to an improved understanding of the ease and speed of human sentence processing. Chapter 6 concludes the book with a short summary.

The book is a scientific work in the best tradition. It lays the groundwork, reports on experiments, presents a model incorporating the experimental results and defends the model against competing proposals. It comes with an extensive reference list. As such the book addresses mostly researchers in a relatively narrow scientific domain, namely people working in cognitive linguistics. I found it demanding to read even with a background in a neighboring discipline. The book is certainly not an introduction to the field nor can it be used as a textbook.

One reason for the efforts required from a thoughtful reader is the pervasive use of acronyms. Here are just two examples:
(p.51) The motivation for the PPP processing input in accord with the strategies LC and MA is that ...

(p.150) Before contrasting MT and IT and motivating a theory of SI parsing, it is necessary ...

Certainly, all but the most obvious acronyms are introduced in the book at one point or another. But they add up to such a number that they result in a cognitive burden. This is regrettable, especially since an acronym glossary is missing.

Other than this, the book has been crafted with great care. Layout and tree structure graphs are excellent and there are very few misspellings. One would wish that more publications were on such a level of formal quality.

Bradley L. Pritchett (1992): Grammatical Competence and Parsing Performance

Carnegie Mellon University.

Book Review: by Gerry Altmann, University of Sussex: PDF

Markus Bader, Michael Meng (1999): Subject-Object Ambiguities in German Embedded Clauses: An Across-the-Board Comparison

Journal of Psycholinguistic Research, Vol.28, No.2. (PDF)

Abstract: This paper examines the processing of embedded clauses in German which are ambiguous between a subject-before-object and an object-before-subject order. In an experiment using a speeded grammaticality judgment task, four types of locally ambiguous clauses were compared: (i) sentences involving movement of a definite noun phrase (NP), (ii) sentences involving pronoun movement, (iii) relative clauses, and (iv) embedded questions. We found that readers were consistently garden-pathed in the object-before-subject condition, regardless of sentence type. Furthermore, there were considerable differences with respect to garden-path strength. The garden-path effect was strongest for sentences involving scrambling. In addition, sentences involving pronoun movement induced more processing difficulty than embedded questions and relative clauses. We argue that our findings can be best explained within a serial processing model that acknowledges both syntactic and nonsyntactic influences on reanalysis and that can account for graded effects of garden-path strength.

Fabrizio Costa, P. Frasconi, V. Lombardo, and G. Soda (2003): Towards Incremental Parsing of Natural Language using Recursive Neural Networks

Applied Intelligence, 19 (1-2):9-25, July - October, 2003, PDF

F. Costa, V. Lombardo, P. Frasconi and G. Soda: Wide coverage incremental parsing by learning attachment preferences

Conference of the Italian Association for Artificial Intelligence, 2001, PDF

Gardent Claire and Bonnie Webber (1998): Describing Discourse Semantics.

Proceedings of the 4th TAG+ workshop, Philadelphia, US, 1998. (PS, citeseer)

(move this paper into another category)
 
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback