Search

MUSE is a computer model for natural language processing, based on a semantic memory network like that of Quillian's TLC. MUSE, from a Model to Understand Simple English, processes English sentences of unrestricted content but somewhat restricted format. The model first applies syntactic analysis to eliminate some interpretations and then employs a simplified semantic intersection procedure to find a valid interpretation of the input. While the semantic processing is similar to TLC's, the syntactic component includes the early use of parse frees and special purpose rules. The "relational triple" notation used during interpretation of input is compatible with MUSE's memory structures, allowing direct verification of familiar concepts and the addition of new ones. MUSE also has a repertoire of actions, which range from editing and reporting the contents of its own memory to an indirect form of question answering. Examples are presented to demonstrate how the model interprets text, resolves ambiguities, adds information to memory, generalizes from examples, and performs various actions. [ABSTRACT FROM AUTHOR]

The use of augmented transition network grammars for the analysis of natural language sentences is described. Structure-building actions associated with the arcs of the grammar network allow for the reordering, restructuring, and copying of constituents necessary to produce deep-structure representations of the type normally obtained from a transformational analysis, and conditions on the arcs allow for a powerful selectivity which can rule out meaningless analyses and take advantage of semantic information to guide the parsing. The advantages of this model for natural language analysis are discussed in detail and illustrated by examples. An implementation of an experimental parsing system for transition network grammars is briefly described. [ABSTRACT FROM AUTHOR]

A practical PL/1 program is described which can assist comparative linguists to determine the regular sound correspondences between genetically related languages. The investigator must arrange data for input by aligning pairs of suspected cognates. The program tabulates the correspondences, and uses list processing techniques to sort and count them. Each pair of words is then assigned a relative value that is a function of the total frequency in the data of each correspondence found in that pair of words. The output is a list of all correspondence types with their frequency of occurrence in the data, and a separate listing of each correspondence with all words pairs showing that correspondence (unless their relative value is below an arbitrarily chosen cutoff point). The article explains the usefulness, as well as the limitations, of the program, and illustrates its use with a small portion of hypothetical data. [ABSTRACT FROM AUTHOR]

Although automatic syntactic and semantic analysis is not yet possible for all of an unrestricted natural language text, some applications, of which content analysis is one, do not have such a stringent coverage requirement. Preliminary studies show that the Harvard Syntactic Analyzer can produce correct and unambiguous identification of the subject and object of certain verbs for approximately half of the relevant occurrences. This provides a degree of coverage for content analysis variables which compares favorably to manual methods, in which only a sample of the total available text is normally processed. [ABSTRACT FROM AUTHOR]

Recent experiments in programming natural language question-answering systems are reviewed to summarize the methods that have been developed for syntactic, semantic, and logical analysis of English strings. It is concluded that at least minimally effective techniques have been devised for answering questions from natural language subsets in small scale experimental systems and that a useful paradigm has evolved to guide research efforts in the field. Current approaches to semantic analysis and logical inference are seen to be effective beginnings but of questionable generality with respect either to subtle aspects of meaning or to applications over large subsets of English. Generalizing from current small-scale experiments to language-processing systems based on dictionaries with thousands of entries-with correspondingly large grammars and semantic systems-may entail a new order of complexity and require the invention and development of entirely different approaches to semantic analysis and question answering. [ABSTRACT FROM AUTHOR]

The Teachable Language Comprehender (TLC) is a program designed to be capable of being taught to "comprehend" English text. When text which the program has not seen before is input to it, it comprehends that text by correctly relating each (explicit or implicit) assertion of the new text to a large memory. This memory is a "semantic network" representing factual assertions about the world. The program also creates copies of the parts of its memory which hove been found to relate to the new text, adapting and combining these copies to represent the meaning of the new text. By this means, the meaning of all text the program successfully comprehends is encoded into the same format as that of the memory. In this form it can be added into the memory. Both factual assertions for the memory and the capabilities for correctly relating text to the memory's prior content are to be taught to the program as they are needed. TLC presently contains a relatively small number of examples of such assertions and capabilities, but within the system, notations for expressing either of these are provided. Thus the program now corresponds to general process for comprehending language, and it provides methodology for adding the additional information this process requires to actually comprehend text of any particular kind. The memory structure and comprehension process of TLC allow new factual assertions and capabilities for relating text to such stored assertions to generalize automatically. That is, once such an assertion or capability is put into the system, it becomes available to help comprehend a great many other sentences in the future. Thus the addition of a single factual assertion or linguistic capability will often provide large increment in TLC's effective knowledge of the world and in its overall ability to comprehend text. The program's strategy is presented as a general theory of language comprehension. [ABSTRACT FROM AUTHOR]

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. [ABSTRACT FROM AUTHOR]

Easy English is a natural command language designed to simplify communication between man and machine through remote typewriter console. It has been developed for retrieval of documents from a computerized data base, the Moore School Information Systems Laboratory files. Requests are formulated in a standardized syntactical form (examples of which are presented), and this form is then transformed into an equivalent query expressed in the retrieval system's original Symbolic Command Language, which is briefly described. Operation of Easy English is detailed by illustration of the transformations performed upon a sample request up to the point at which the request string is sent to the system. A macro flowchart of Easy English is included, and an Appendix provides the printout of a retrieval demonstration. [ABSTRACT FROM AUTHOR]

A further development of a computer program (ELIZA) capable of conversing in natural language is discussed. The importance of context to both human and machine understanding is stressed. It is argued that the adequacy of the level of understanding achieved in a particular conversation depends on the purpose of that conversation, and that absolute understanding on the part of either humans or machines is impossible. [ABSTRACT FROM AUTHOR]

The article discusses the emergence of natural language translation by computer. There are only two known private companies in the United States translating natural languages by computer, one of them is Logos Development Corporation of Middletown, New York. Logos is under private contract to translate English manuals into Russian for a large truck plant to be built by American firms in the Soviet Union. The other known commercial firm in machine translation is Latsec, Inc. in La Jolla, California. Latsec developed the Systran system for Foreign Technology Division of Wright-Patterson Air Force Base.

The article focuses on the formation of the Association for Computing Machinery's (ACM) Special Interest Committee on Language Analysis and Studies in the Humanities (SICLASH). In recognition of the need for closer rapport among scholars and scientists engaged in computer-aided language analysis, computer-aided humanistic studies, and the development of software and hardware for such research, the Association for Computing Machinery has chartered the SICLASH. The goals of the Committee are to provide a focus within the ACM for a broad spectrum of computer-aided language studies. Various professional organizations have relevant interests in particular aspects of such research, but there is no single organization permitting people from disparate applications groups to share information and research on common problems in computer-based language analysis. It also provides a focus within the ACM for those humanists within the fine arts, such as music and graphic arts, who have need of computer technology, as well as for the majority of Humanists concerned with the written and spoken language.