Janet Dean Fodor

The sausage machine: A new two-stage parsing model

Abstract It is proposed that the human sentence parsing device assigns phrase structure to word strings in two steps. The first stage parser assigns lexical and phrasal nodes to substrings of roughly six words. The second stage parser then adds higher nodes to link these phrasal packages together into a complete phrase marker. This model of the parser is compared with ATN models, and with the two-stage models of Kimball (1973) and Fodor, Bever and Garrett (1974). Our assumption that the units which are shunted from the first stage to the second stage are defined by their length, rather than by their syntactic type, explains the effects of constituent length on perceptual complexity in center embedded sentences and in sentences of the kind that fall under Kimballs principle of Right Association. The particular division of labor between the two parsing units allows us to explain, without appeal to any ad hoc parsing strategies, why the parser makes certain ‘shortsighted’ errors even though, in general, it is able to make intelligent use of all the information that is available to it.

REFERENTIAL AND QUANTIFICATIONAL INDEFINITES

ConclusionThe formal semantics that we have proposed for definite and indefinite descriptions analyzes them both as variable-binding operators and as referring terms. It is the referential analysis which makes it possible to account for the facts outlined in Section 2, e.g. for the purely ‘instrumental’ role of the descriptive content; for the appearance of unusually wide scope readings relative to other quantifiers, higher predicates, and island boundaries; for the fact that the island-escaping readings are always equivalent to maximally wide scope quantifiers; and for the appearance of violations of the identity conditions on variables in deleted constituents. We would emphasize that this is not a random collection of observations. They cohere naturally with each other, and with facts about other phrases that are unambigously referential.We conceded at the outset of this paper that the referential use of an indefinite noun phrase does not, by itself, motivate the postulation of a referential interpretation. Our argument has been that the behavior of indefinites in complex sentences cannot be economically described, and certainly cannot be explained, unless a referential interpretation is assumed. It could be accounted for in pragmatic terms only if the whole theory of scope relations and of conditions on deletion could be eliminated from the semantics and incorporated into a purely pragmatic theory. But this seems unlikely.

Learning To Parse

A strong claim about human sentence comprehension is that the processing mechanism is fully innate and applies differently to different languages only to the extent that their grammars differ. If so, there is hope for an explanatory project which attributes all parsing “strategies” to fundamental design characteristics of the parsing device. However, the whole explanatory program is in peril because of the discovery (Cuetos & Mitchell, 1988) that Late Closure is not universal: Spanish, and also Dutch and other languages, favor Early Closure (high attachment) where English favors Late Closure flow attachment). I argue that the universal parser can weather this storm. Exceptions to Late Closure in Spanish and other languages are observed only in one construction (a relative clause attaching into a complex noun phrase [NP]), which is borderline in English too. For other constructions, low attachment is preferred in all languages tested. I propose that what differentiates the complex NP construction is the heaviness of the attachee compared to that of the host configuration. A relative clause is a heavy attachee, and the lower NP alone is small as a host; the relative is therefore better balanced if the whole complex NP is its host. A wide range of facts is accounted for by the principle that a constituent likes to have a sister of its own size. Light constituents will tend to attach low, and heavy ones to attach high, since larger constituents are dominated by higher nodes. A preference for balanced weight is familiar from work on prosodic phrasing. I suggest, therefore, that prosodic processing occurs in parallel with syntactic processing (even in reading) and influences structural ambiguity resolution. Height of attachment ambiguities are resolved by the prosodically motivated same-size-sister constraint. The exceptional behavior of English may be due to its prosodic packaging of a relative pronoun with the adjacent noun, overriding the balance tendency. If this explanation is correct, it is possible that all cross-language variations in parsing preferences are due to cross-language variations in the prosodic component of the competence grammar.

Reanalysis in sentence processing

Preface. 1. Prosodic Influences on Reading Syntactically Ambiguous Sentences M. Bader. 2. Reanalysis Aspects of Movements M. de Vincenzi. 3. Syntactic Reanalysis, Thematic Processing, and Sentence Comprehension F. Ferreira, J.M. Henderson. 4. Attach Anyway J.D. Fodor, A. Inoue. 5. Sentence Reanalysis, and Visibility L. Frazier, C. Clifton, Jr. 6. Diagnosis and Reanalysis: Two Processing Aspects the Brain May Differentiate A.D. Friederici. 7. Syntactic Analysis and Reanalysis in Sentence Processing P. Gorrell. 8. Reanalysis and Limited Repair Parsing: Leaping off the Garden Path R.L. Lewis. 9. A Computational Model of Recovery V. Lombardo. 10. Parsing as Incremental Restructuring S. Stevenson. 11. Generalized Monotonicity for Reanalysis Models P. Sturt, M.W. Crocker.

The diagnosis and cure of garden paths

We propose that, for the human parser, recovery from garden paths consists in repairing the structure built so far, rather than reparsing the input. The difficulty of a repair is attributable not to the cost of effecting the structural alterations but to the cost of deducing which alterations are needed. The parser must diagnose its error in order to correct it. The error is signaled by an input word that is incompatible with the current structure; this is the symptom from which the diagnosis must be made. If the error is transparently clear from the nature of the symptom, recovery is easy; but sometimes the necessary reasoning is obscure, and then the diagnosis is unsuccessful and the garden path persists. Unlike other repair models, the diagnosis model needs no special mechanism for revising garden path analyses. The garden path recovery device is the same machine as the first-pass parser, merely set into emergency mode. When faced with a breakdown the parser does not stop its normal activities and enter a new mode of reasoning to detect what went wrong. It simply continues to parse, attaching the problematic input item in the least ungrammatical way it can, despite the conflict with previously built structure. This conflict is productive; it provokes adjustments to the existing structure. In successful cases, one adjustment leads to another until a stable state is reached, at which point the original error will have been eliminated. Examples suggest that the parser gives more weight to syntatctic than to pragmatic acceptability; only a syntactic clash between the input and the existing structure sets the adjustment process in motion.

Parsing to Learn

Learning a language by parameter setting is almost certainly less onerous than composing a grammar from scratch. But recent computational modeling of how parameters are set has shown that it is not at all the simple mechanical process sometimes imagined. Sentences must be parsed to discover the properties that select between parameter values. But the sentences that drive learning cannot be parsed with the learners current grammar. And there is not much point in parsing them with just one new grammar. They must apparently be parsed with all possible grammars, in order to find out which one is most successful at licensing the language. The research task is to reconcile this with the fact that the human sentence parsing mechanism, even in adults, has only very limited parallel parsing capacity. I have proposed that all possible grammars can be folded into one, if parameter values are fragments of sentential tree structures that the parser can make use of where necessary to assign a structure to an input sentence. However, the problem of capacity limitations remains. The combined grammar will afford multiple analyses for some sentences, too many to be computed on-line. I propose that the parser computes only one analysis per sentence but can detect ambiguity, and that the learner makes use of unambiguous input only. This provides secure information but relatively little of it, particularly at early stages of learning where few grammars have been excluded and ambiguity is rife. I consider three solutions: improving the parsers ability to extract unambiguous information from partially ambiguous sentences, assuming default parameter values to temporarily eliminate ambiguity, reconfiguring the parameters so that some are subordinate to others and do not present themselves to the learner until the others have been set. A more radical alternative is to give up the quest for error-free learning and permit parameters to be set without regard for whether the parser may have overlooked an alternative analysis of the sentence. If it can be assumed that the human parser keeps a running tally of the parameter values it has accessed, then the learner would do nothing other than parse sentences for comprehension, as adults do. The most useful parameter values would become more and more easily accessed; the noncontributors would drop out of the running. There would be no learning mechanism at all, over and above the parser. But how accurate this system would be remains to be established.

Empty categories in sentence processing

Abstract An empty category is an inaudible/invisible constituent of a sentence, postulated by linguists to account for certain regularities of sentence structure. To identify an empty category and associate it with an antecedent that will determine its interpretation, a sentence processing device must apply considerable linguistic knowledge, both universal and language-particular. Experiments show that human processing of empty category constructions is efficient and linguistically informed. Different linguistic theories postulate different empty categories. Recent and ongoing experiments by several researchers attempt to determine which kinds of empty categories are recognised by the human processor. Assuming a transparent relationship between the parser and the grammar, the results of these studies could contribute to the evaluation of competing linguistic theories.

Using cross-modal lexical decision tasks to investigate sentence processing

Recent investigations of sentence processing have used the cross-modal lexical decision task to show that the antecedent of a phonologically empty noun phrase (specifically, WH-trace) is reactivated at the trace position. G. McKoon, R. Ratcliff, and G. Ward (1994) claimed that (a) a design feature concerning the choice of related and unrelated targets is a possible confound in this work and (b) the conclusions drawn from this previous research are therefore called into question. These claims are considered in light of both McKoon et al.s experimental findings and results of our own experiments in which we test their linguistic materials. We argue that their results may be due to the nature of their materials. Additionally, we argue that a follow-up experiment reported by G. McKoon and R. Ratcliff (1994) used a technique that is not comparable to the cross-modal lexical decision task. It is concluded that current evidence supports the claim that structural information is using during on-line sentence processing and that the cross-modal technique is sensitive to this.

Sentence matching and overgeneration

Abstract Freedman and Forster (1985) claim that sentence matching times for ungrammatical sentences demonstrate the psychological reality of different types of ungrammaticality, and that this implies that sentences are mentally assigned multilevel syntactic derivations as characterized by Government Binding Theory. We question the notion of overgeneration which links F&Fs conclusions to their data. And we present further experimental results which suggest that the observed differences among ungrammatical sentences do not reflect their linguistic status. Rather, they are due to a tendency for subjects to spontaneously correct some, but not all, kinds of ungrammaticality.

Contextual information and temporal terms.

We report an experiment designed to identify how contextual information can influence childrens performance on an experimental task involving temporal terms. Crain (1982) reported improved performance on a comprehension task when subjects were provided with contextual information, and he suggested that the improvement was due to satisfaction of presuppositions. However, this contextual information might have served to simplify task demands by providing prior information concerning an important aspect of the task. The present study distinguishes these factors by incorporating contextual information into the subordinate clause of the test sentences in a comprehension experiment (to satisfy presupposions) or into the main clause (to provide comparable prior information without satisfying presuppositions). We conclude that contextual information results in a significant improvement only when such information can be used to satisfy presuppositions.