Showing papers on "Rule-based machine translation published in 1969"

Notes on top-down languages

Abstract: Properties of context-free languages and grammars permitting deterministic top-down recognition with bounded lookahead are discussed. In particular, it is shown that for eachk>1 there are such languages requiring a lookahead of at leastk characters.

Property grammars and table machines

Abstract: A method is given for generalizing context free grammars to include infinite terminal sets. The new grammar has many of the properties of context free grammars (e.g. membership is decidable and recognition by a suitably generalized nondeterministic pushdown device is always possible). Yet, treating each identifier as a separate terminal symbol, the grammar has enough power to describe all of the syntactic and semantic constraints of the declaration structure for variables and labels in ALGOL-60.

Directed random generation of sentences

Abstract: The problem of producing sentences of a transformational grammar by using a random generator to create phrase structure trees for input to the lexical insertion and transformational phases is discussed. A purely random generator will produce base trees which will be blocked by the transformations, and which are frequently too long to be of practical interest. A solution is offered in the form of a computer program which allows the user to constrain and direct the generation by the simple but powerful device of restricted subtrees. The program is a directed random generator which accepts as input a subtree with restrictions and produces around it a tree which satisfies the restrictions and is ready for the next phase of the grammar. The underlying linguistic model is that of Noam Chomsky, as presented in Aspects of the Theory of Syntax. The program is written in FORTRAN IV for the IBM 360/67 and is part of a unified computer system for transformational grammar. It is currently being used with several partial grammars of English.

W-grammar

Abstract: A new type of grammars is presented here, called W-grammars. It is shown how they can be used in translation processes. Examples are taken from the fields of algebraic manipulation and computational linguistics.

Properties of formal grammars with mixed types of rules and their linguistic relevance

Abstract: In this paper, we will study a class of formal grammars with mixed types of rules. The reason for considering such grammars is that no single style (i.e., formal character of rules) of formal grammars is able to represent the various aspects of language structure in a natural way. Various considerations for setting up such grammars have been discussed. Generation schemes which map strings in the language of one mixed grammar into strings in the language of another mixed grammar (both strings being 'well-formed') have been studied. Linguistic relevance of these concepts has also been discussed.

Linguistic specification and analysis of classes of line patterns.

Abstract: : Pictorial patterns can be considered as statements in a two-dimensional language. In the report, a variety of languages composed of sets of encoded geometric curves are classified in the Chomsky language hierarchy. The relationships between classes of languages and classes of automata then permit bounds on the time and memory required to recognize the patterns to be determined. The phrase-structure grammar scheme for string languages is extended by permitting symbols to have an arbitrary number of 'attaching points,' and classes of the extended languages are defined. An algorithm for parsing a 'context-free' language of this type according to a formal language description is incorporated into a table-driven pattern analyzer. This device is an extension of the idea of a table-driven compiler and is able to perform a broad range of pattern analysis tasks on widely differing classes of line patterns. The class of patterns and method for structurally breaking down the patterns can be changed simply by changing the entries in the program syntax tables. The table-driven pattern analyzer is applied to the recognition of events in bubble chamber photographs. (Author)

A system for designing fast programming language translators

Abstract: This paper demonstrates a straightforward algorithm for converting programming language grammars into pushdown-store automata translators. The language grammar is written as a "translation grammar" in which, for each syntactic rule, there is a corresponding "rule of translation" that recursively specifies the reverse Polish string translation of the objects in the syntactic rule. This augmented grammar is transformed directly into a flow chart for the appropriate translator.

Semantics and the syntactic classification of words

Abstract: Traditional grammars classify words according to generic syntactic functions or morphological characteristics. For teaching humans and for descriptive linguistics this seemed sufficient. The advent of computers has changed the situation. Since machines are devoid of experiential knowledge, they need a more explicit grammar to handle natural language. Correlational Grammar is an attempt in that direction. The paper describes parts of correlational syntax and shows how a highly differentiated syntax can be used to establish word classes for which an intensional semantic definition can then be found. It examplifies this approach in two ares of grammar: predicative adjectives and transitive verbs. The classification serves to eliminate ambiguity and spurious computer interpretations of natural language sentences.

Translation model with semantic capability

Abstract: Automatic translation systems may be characterized for the most part in one of two ways: (1) Translators in which the language processors are internally dependent on the description of the syntax of the language being processed. (2) Translators externally dependent on the syntax.* Programming a translation system of the first type has two major advantages. It can be set up quicker and it requires less initial technical investment. On the other hand, such a system imposes many operational difficulties on debugging the linguistic description. The second type is better adapted to more flexible linguistics research but is more costly and time-consuming to establish since its concrete results and more sophisticated research power and processing capabilities are not realized for some time after the initial effort. The linguistic descriptive effort which has gone into either of these approaches has tended to concentrate primarily on matters of morphosyntactic analysis, rearrangement and synthesis. Such efforts as have been directed to semantic description in operational translation systems have been limited virtually to ad hoc statements of co-occurrence of particular items. Little else could have been done due, perhaps, to the lack of an adequate theoretic and formalizable notion of semantic process and to the descriptive limitations inherent in processing algorithms developed thus far. In this paper I present an outline of the Linguistics Research System, a stratified linguistic data processing system, and interpret some of its facilities for application to semantic description. The system is of the second type in that the linguistic description is in the form of a phrase

Grammars and Languages

Abstract: (1969). Grammars and Languages. The American Mathematical Monthly: Vol. 76, No. 4, pp. 335-342.

The algebra of derivations and a semithue parser

Abstract: While derivations are used to formally generate sentences from the productions of a grammar, the structure, or structural descriptions, which the grammar imposes on a sentence is the entity of general interest. For context free generation, a tree is an adequate representation of sentence structure. For more general classes of grammars, a tree is not adequate. With the strong form of derivation developed in Griffiths3 and Benson1 an algebra of derivations is possible. In particular there is an equivalence relation on derivations such that the equivalence classes are a representation of sentence structure for semithue languages. A slightly different equivalence relation provides sentence structure for context sensitive grammars.There is a parser for context free and semithue languages which is closely allied to the equivalence class representation of sentence structure. The parser is exemplified, described, and shown to find each structure assigned to the input sentence exactly once.

A Note on Two Basic Semantic Distinctions.

Abstract: : The paper discusses the nature of two basic semantic distinctions--intensional/extensional, and mental/physical (mental/physical being similar to abstract/concrete but more explicit)--and how aan understanding of their interaction is an essential preliminary to writing a semantically motivated grammar of English. (Author)

Syntactical decoders and backtracking s-grammars,

Abstract: : It is proposed that syntactical information be used to aid in the post-information-theoretic decoding of formal languages that have been transmitted through a noisy channel A model for an information source producing a continuous stream of grammatical sentences is exhibited General desirable properties of syntactical decoders are proposed, and a general minimum-distance decoding algorithm established Backtracking 5-grammars are defined, and used to exemplify backtracking algorithms for attempted correction of single and multiple errors (Author)

A pragmatic approach to machine translation from Chinese to English

Abstract: Chinese machine translatlon can be achieved by organizing all the necessary llng~istic data in such a way that the computer can compare and retrieve them in the most economical way. We are constantly reminded that the storage space in the co~uter is limited, and the processing time is expensive. We mmst aim at the efficiency of the system without sacrificing accuracy. Five types of linguistic data have to be stored in the computer before a translation can be rendered: (1) a Chinese to English dictionary, (2) Chinese syntactic rules, (3) syntactic converslon rules from Chinese structure to English structure, (4) English morphological rules, and (5) the text to be translated. (1) The dictionary must have the capability to distinguish automatically the different meanings which a Chinese lexeme represents. l~anmn readers can do this and 8o can the machine if enough linguistic information is stored in the computer. To a certain extent the meanings of a given lexema can be differentiated by its different syntactic functions. For example the word i means 'one' when it 18 used as a numeral. A numeral i8 defined as the class of words which is followed by a classifier, a measure, a collective and a

On coordination reduction and sentence analysis

Abstract: A class of coordination phenomena in natural languages is considered within the framework of transformational theory. To account for these phenomena it is proposed that certain machinery be added to the syntactic component of a transformational grammar. This machinery includes certain rule schemata, the conditions under which they are to be applied, and conditions determining the sequence of subtrees on which they are to be performed. A solution to the syntactic analysis problem for this class of grammars is outlined. Precise specification of both the generative procedure of this paper and its inverse is given in the form of LISP function definitions.