Showing papers on "Phrase published in 1985"

System for translation of sentences from one language to another

Abstract: A machine translation system for translating a sentence of a first language to a sentence of a second language. The system has a dictionary, a display, an input unit, a selection data memory and a combination data memory. The translation possibilities for respective words and/or phrases of the original are obtained from the dictionary, ordered and are displayed by the display. The operator operates the input unit, thereby selecting one of the translation possibilities for each word or phrase. The selected possibilities for the words and phrases form a sentence of the second language, and are stored in the data memory in such a manner that they are associated with the equivalents of the first language. Further, the selected possibilities are stored in the combination data memory in such a way that any selected words or phrases that are semantically related are associated with the first language equivalent to one of the related words or phrases. The translation possibilities for any repeatedly used word or phrase of the first language are display in the priority order defined by the data stored in data memories.

Figures of speech

Abstract: An educational model for teaching language grammar, parts of speech, syntax, and sentence structure is disclosed in which thirty (30) individual elements of speech are identified and grouped into ten (10) distinct sets which are embodied in physical models. Each model represents a word or group of words, referred to as "elements," which form the logical building blocks of the language. Certain elements are actually hybrid elements in that they reflect the unique properties and derivations of verbals, phrases and clauses. The models interengage with one another without fully interlocking, in order to designate, simultaneously, each element's part of speech and part of sentence within the sentence. Each model is generic in that it does not impose case or agreement of person and number or predetermined word choices on the word or group of words it designates. Case is determined only by the position of each element as it interengages with other elements. Symbols are used to designate the function of each element; color is used to designate the element's set, and shape is used to designate sentence part and part of speech. The user interengages each element with other elements to form a phrase, clause, or sentence as an alternative to conventional sentence diagramming. The models may be embodied in print, plastic, video, or software.

Syntax, preference, and right attachment

Abstract: The paper claims that the right attachment rules for phrases originally suggested by Frazier and Fodor are wrong, and that none of the subsequent patchings of the rules by syntactic methods have improved the situation. For each rule there are perfectly straightforward and indefinitely large classes of simple counterexamples. We then examine suggestions by Ford et al., Schubert and Hirst which are quasi-semantic in nature and which we consider ingenious but unsatisfactory. We offer a straightforward solution within the framework of preference semantics, and argue that the principal issue is not the type and nature of information required to get appropriate phrase attachments, but the issue of where to store the information and with what processes to apply it. We present a prolog implementation of a best first algorithm covering the data and contrast it with closely related ones, all of which are based on the preferences of nouns and prepositions, as well as verbs.

Multiple subcategorization and the ϑ-criterion: The case of climb

Abstract: We have successfully accounted for all the data of section 2 with the single unified entry (27). The price has been two innovations in the semantic structure of lexical items: the curly bracket notation for alternative realizations of variables, and the P operator. Each of these has been independently motivated. Returning to the general issue that motivated this study, we have maintained a strong correspondence between syntactic ϑ-positions and semantic argument positions. However, the correspondence need not be strictly one-to-one. We have seen multiple argument positions for a single ϑ-position in transitive climb, transaction verbs such as buy and sell, and possibly intransitive dress. We have also seen, possibly, that an argument may be multiply filled, once by a subcategorized phrase and once by a nonsubcategorized phrase. We have also seen that the correspondence between syntactic and semantic structure is encoded in lexical entries by means of principles more complex than seems to have been suspected. And the complexities we have found are for climb, which intuition suggests is a relatively simple item with rather transparent semantics. When we attempt to represent verbs with complicated options for sentential complementation (such as know and ask), we should expect the descriptive problems to multiply. Eventually, of course, one would like to adequately constrain the theory of syntax-semantics linkages in the lexicon. The present study should caution us, however, of the danger of applying Occam's razor too soon, thereby cutting off one's hand. This study has also provided evidence confirming Grimshaw's (1979) conclusion that syntactic and semantic conditions on lexical insertion are autonomous. Here we have ended with a lexical entry whose sole syntactic constraint is that there is a single postverbal complement. The syntactic category of the complement, however, is determined by the range of semantic categories possible in the corresponding variables. Viewed from a different angle, though, one might say that it is the availability of only one syntactic complement position that prevents both semantic variables from being expressed at once. There is nothing conceptually wrong with * Bill climbed the mountain up the ropes, in which both postverbal complements fill j-variables in (27) — it is just a syntactic fact about the English verb climb that makes it ungrammatical. To sum up, then, the subcategorization feature is not simply a projection of argument structure, as generally assumed. Rather, ϑ-structure and argument structure are better thought of as each constraining potential projections of the other. Finally, it is evident that these issues could not have been investigated without a fairly explicit theory of semantic structure like Conceptual Semantics. I hope to have provided here a taste of what syntactic theory may have missed out on, as a result of its habit of viewing semantics through the narrow window of more traditional formalizations of predicate-argument structure. I am grateful to Jane Grimshaw and David Olson for helpful discussion of this material, and to Jerrold Katz and especially another (anonymous) NLLT reviewer for many insightful comments on an earlier version. The puzzle of rent in section 4 goes back to discussions with Joe Emonds around 1966. This research was supported in part by NSF Grant IST-8120403 to Brandeis University, and in part by NSF Grant BNS-7622943 to the Center for Advanced Study in the Behavioral Sciences, where my initial exploration of climb took place one typically glorious day in February 1984.

A Structure-Sharing Representation for Unification-Based Grammar Formalisms

Abstract: This paper describes a structure-sharing method for the representation of complex phrase types in a parser for PATR-II, a unification-based grammar formalism.In parsers for unification-based grammar formalisms, complex phrase types are derived by incremental refinement of the phrase types defined in grammar rules and lexical entries. In a naive implementation, a new phrase type is built by copying older ones and then combining the copies according to the constraints stated in a grammar rule. The structure-sharing method was designed to eliminate most such copying; indeed, practical tests suggest that the use of this technique reduces parsing time by as much as 60%.The present work is inspired by the structure-sharing method for theorem proving introduced by Boyer and Moore and on the variant of it that is used in some Prolog implementations.

Perception of stress pattern and word recognition

Abstract: Theories of spoken word recognition do not explicitly take account of nonsegmental properties of the stimulus. The present research investigates how, and to what extent, acoustic information on the stress pattern of a (Dutch) word contributes to word recognition, both in hifi degraded natural speech and in synthetic speech (concatenated diphones). In a series of experiments, words with correct or incorrect stress position in the first, second, or third syllable were presented in isolation or preceded by a short carrier phrase. Words were made audible in fragments of increasing size (“gating”) so as to trace the respective contribution of stressed and unstressed syllables to the word isolation process as the stimulus develops in time. Whole word recognition was included in some of the experiments as a control task. Selected results will be presented and some of the consequences for word recognition models will be discussed.

Repetition in schizophrenic speech.

Abstract: Previous research on language with the Type-Token Ratio (TTR) indicates an increased repetitiousness in schizophrenics, specifically in thought-disordered schizophrenics. The present investigation involves an analysis of oral language samples utilizing a variety of repetition measures. Subjects were 40 schizophrenics, 20 affective disorder controls and 22 normal controls. Phrase repetitions were found to be more prevalent among the schizophrenic group, and especially among the TD subgroup, and indices of phrase repetition were positively correlated with dimensions of formal thought disorder. The frequency patterns of repeated words in schizophrenic language reflect a less common, but not unusual lexicon compared to controls, thereby supporting the view that word repetition in schizophrenic utterance is not a consequence of a restricted vocabulary range. The intervals between phrase repetitions were not different for schizophrenics and controls, indicating that simple perseveration is also not a likely exp...

Aspects of Pronoun Interpretation: Evaluation of Search Formulations of Inference

Abstract: Some recent experimentation shows considerable differences in the mechanisms underlying noun - phrase and pronoun reference resolution using reading - time and other procedures The arguments will revolve round these differences, and will concentrate upon some general characteristics of the procedures associated with pronouns One of the major side effects is to differentiate a number of options which have hitherto been lumped together as inferential processes

Interaction of stimulus and contextual information during reading: Identifying words within sentences.

Abstract: In general, studies on the effects of a sentence context on word identification have focused on how context affects the efficiency of processing a single target word, presented separately from the context. Such studies probably would be incapable of measuring contextual facilitation resulting from cascaded or parallel processing of neighboring words within a sentence. To measure these and other types of facilitation, we presented entire phrases and sentences for subjects to read as fast as possible and to monitor for nonwords. Subjects read at rates representative of natural reading. Experiment 1 demonstrated a large contextual facilitation effect on decision time. Experiment 2 showed that facilitation is caused by specific semantic information and, perhaps to a greater degree, by nonpredictive syntactic information. Experiment 3 showed that the amount of facilitation is greater than could be accounted for by separate contributions from autonomous word level and sentence level processes. These results present difficulties for an autonomous model of reading, but are consistent with interactive models, in which the results of ongoing sentential analyses are combined with stimulus information to identify words.

New approaches to parsing conjunctions using prolog

Abstract: Conjunctions are particularly difficult to parse in traditional, phrase-based grammars. This paper shows how a different representation, not based on tree structures, markedly improves the parsing problem for conjunctions. It modifies the union of phrase marker model proposed by Goodall [1984], where conjunction is considered as the linearization of a three-dimensional union of a non-tree based phrase marker representation. A PROLOG grammar for conjunctions using this new approach is given. It is far simpler and more transparent than a recent phrase-based extraposition parser conjunctions by Dahl and Mc.Cord [1984]. Unlike the Dahl and Mc.Cord or ATN SYSCONJ approach, no special trail machinery is needed for conjunction, beyond that required for analyzing simple sentences. While of comparable efficiency, the new approach unifies under a single analysis a host of related constructions: respectively sentences, right node raising, or gapping. Another advantage is that it is also completely reversible (without cuts), and therefore can be used to generate sentences.

Intonation in text‐to‐speech synthesis: Evaluation of algorithms

Abstract: Two algorithms, termed schematic and naturalistic, for generating intonation contours in an English text‐to‐speech system are compared by eliciting preference judgments from a total of 21 subjects. The major problem for both algorithms, but especially for the schematic algorithm, has to do with accent assignment and with the determination of the intonation phrase rather than with the phonetic realization of accent through manipulation of F0. Due to parser errors, phrase boundaries are incorrectly identified in 30% of the sentences used in the three experiments. Moreover, the naturalistic algorithm uses a grammatical part‐of‐speech hierarchy which ranks nouns higher than verbs. Therefore, incorrect classification of verbs as nouns (the major classification error) results in an unintended accent. The results indicate that accent assignment and phrase determination are the primary areas requiring improvement in order to further increase the naturalness of synthetic speech intonation.

Towards a Self-Extending Lexicon

Abstract: The problem of manually modifying the lexicon appears with any natural language processing program. Ideally, a program should be able to acquire new lexical entries from context, the way people learn. We address the problem of acquiring entire phrases, specifically figurative phrases, through augmenting a phrasal lexicon. Facilitating such a self-extending lexicon involves (a) disambiguation---selection of the intended phrase from a set of matching phrases, (b) robust parsing---comprehension of partially-matching phrases, and (c) error analysis---use of errors in forming hypotheses about new phrases. We have designed and implemented a program called RINA which uses demons to implement functional-grammar principles. RINA receives new figurative phrases in context and through the application of a sequence of failure-driven rules, creates and refines both the patterns and the concepts which hold syntactic and semantic information about phrases.

Three theories of nominalized predicates

Abstract: By the term ‘nominalization’ I mean any process which transforms a predicate or predicate phrase into a noun or noun phrase, e.g. ‘feminine’ is transformed into ‘feminity’. I call these derivative nouns abstract singular terms. Our aim is to provide a model-theoretic interpretation for a formal language which admits the occurrence of such abstract singular terms.

Voice rule synthesization system

Abstract: PURPOSE: To obtain natural intonation by providing a means which generates a time change pattern having a fundamental frequency and setting the magnitude of a command of phrase control based on the time length of a voice pause section just before a phrase. CONSTITUTION: An input text is analyzed into individual words by a morpheme analysis means of a Japanses language analysis part 1 to examine their parts of speech. The accent type of each word or phrase is determined by a voice processing part 2. The magnitude of the phrase control command is supplied from a control parameter of standard intonation, which is determined by the pause time just before the phrase, by a control parameter generation part 3 to generate a pitch pattern and a phoneme parameter time series. The generated fundamental frequency and phoneme parameter are sent to a voice synthesis part 4 to output a voice waveform. Thus, natural intonation like speaking is obtained.

Right attachment and preference semantics

Abstract: The paper claims that the right attachment rules for phrases originally suggested by Frazier and Fodor are wrong, and that none of the subsequent patchings of the rules by syntactic methods have improved the situation. For each rule there are perfectly straightforward and indefinitely large classes of simple counter-examples. We then examine suggestions by Ford et al., Schubert and Hirst which are quasi-semantic in nature and which we consider ingenious but unsatisfactory. We point towards a straightforward solution within the framework of preference semantics, set out in detail elsewhere, and argue that the principal issue is not the type and nature of information required to get appropriate phrase attachments, but the issue of where to store the information and with what processes to apply it.

The Good-Enough Family:

Abstract: Donald W. Winnicott coined the phrase “good-enough” to describe the infant's relation to its environment The term is here extended to a discussion of the “good-enough” family in which the existenti...

The Subject-Matter of Ethics

Abstract: Ethics may be described as the theoretical treatment of moral phenomena. I use the phrase “moral phenomena” to cover all those facts, and only those, in describing which we have to use, in a specifically moral sense, such words as “ought”, “right”, “good” and their opposites, or any others which are merely verbal translations of them. (This is not intended as a definition; if it were it, would be circular; for I have had to introduce the phrase “in a specifically moral sense” into my description of moral phenomena.) I have had to do this, because words like “ought”, “right”, and “good” are also used in various non-moral senses, and then Ethics is not directly concerned with the facts which they describe.

Method for controlling display and calibration of analyzed result of sentence structure in natural language processor

Abstract: PURPOSE: To display the analyzed results of sentence structure in various stages without using a wide space to be required for the display of the analyzed results by displaying respective phrases and clauses to be analyzed results at least in one stage of sentence structure analysis with an underline. CONSTITUTION: In the sentence structure analyzing processing, a text previously stored in a text memory 5 is recognized in the order of word, phrase, clause, and sentence nodes in accordance with a sentence structure analyzing rule 42 and the parental relation is stored in a node information storing area 44. A calibrating information storing area 45 stores holding information for holding a word string specified at its analyzing stage under the specified analysis and surface information for storing node numbers collected in each stage of calibration. The analyzed results are displayed by processing the display information and the surface information by a picture display program 43 and showning each node with an underline in each stage collecting respective analyzed results in gradually higher order like a word, a phrase and a clause. COPYRIGHT: (C)1986,JPO&Japio

A probabilistic approach to grammatical analysis of written English by computer

Abstract: Work at the Unit for Computer Research on the English Language at the University of Lancaster has been directed towards producing a grammatically annotated version of the Lancaster-Oslo/Bergen (LOB) Corpus of written British English texts as the preliminary stage in developing computer programs and data files for providing a grammatical analysis of unrestricted English text.From 1981--83, a suite of PASCAL programs was devised to automatically produce a single level of grammatical description with one word tag representing the word class or part of speech of each word token in the corpus. Error analysis and subsequent modification to the system resulted in over 96 per cent of word tags being correctly assigned automatically. The remaining 3 to 4 per cent were corrected by human post-editors.Work is now in progress to devise a suite of programs to provide a constituent analysis of the sentences in the corpus. So far, sample sentences have been automatically assigned phrase and clause tags using a probabilistic system similar to word tagging. It is hoped that the entire corpus will eventually be parsed.

Grammatical analysis by computer of the lancaster-oslo/bergen (lob) corpus of british english texts

Abstract: Research has been under way at the unit for Computer Research on the English Language at the University of Lancaster, England, to develop a suite of computer programs which provide a detailed grammatical analysis of the LOB corpus, a collection of about 1 million words of British English texts available in machine readable form.The first phrase of the project, completed in September 1983, produced a grammatically annotated version of the corpus giving a tag showing the word class of each word token. Over 93 per cent of the word tags were correctly selected by using a matrix of tag pair probabilities and this figure was upgraded by a further 3 per cent by retagging problematic strings of words prior to disambiguation and by altering the probability weightings for sequences of three tags. The remaining 3 to 4 per cent were corrected by a human post-editor.The system was originally designed to run in batch mode over the corpus but we have recently modified procedures to run interactively for sample sentences typed in by a user at a terminal. We are currently extending the word tag set and improving the word tagging procedures to further reduce manual intervention. A similar probabilistic system is being developed for phrase and clause tagging.

Machine translating system

Abstract: PURPOSE: To obtain a translation by means of a unified equivalent term in translation for the same word (phrase) by replacing the equivalent of the word (phrase) of an entry word including the same word phrase) only with the alteration or registration of the equivalent of a single word or phrase. CONSTITUTION: A translating part 5 to execute a translating processing, and a translation dictionary 6 to store knowledge information utilized for the translating processing are provided. When the equivalent term in translation of a certain word (phrase) in an original sentence is replaced with the different equivalent from the previous equivalent, the entry word including the word (phrase) corresponding to the part, whose equivalent is different from the former one, is retrieved in the translation dictionary 6, and the equivalent part corresponding to the word (phrase) in the retrieved entry word is replaced with the new equivalent. As a result, by translating the sentences with the use of the translation dictionary, the corresponding equivalent (part) corresponding to the same word (phrase) is unified. COPYRIGHT: (C)1990,JPO&Japio