| { | |
| "paper_id": "1952", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T10:32:12.649847Z" | |
| }, | |
| "title": "THE FIRST CONFERENCE ON MECHANICAL TRANSLATION", | |
| "authors": [ | |
| { | |
| "first": "Erwin", | |
| "middle": [], | |
| "last": "Reifler", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Literature University of Washington", | |
| "location": { | |
| "settlement": "Seattle", | |
| "region": "Wash" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Leon Dostert", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Georgetown University", | |
| "location": { | |
| "settlement": "Washington", | |
| "region": "D. C" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Olaf", | |
| "middle": [], | |
| "last": "Helmer", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Georgetown University", | |
| "location": { | |
| "settlement": "Washington", | |
| "region": "D. C" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Harry", | |
| "middle": [ | |
| "D" | |
| ], | |
| "last": "Huskey", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Rand Corporation", | |
| "location": { | |
| "settlement": "Santa Monica, Calif" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Mr", | |
| "middle": [ | |
| "Duncan" | |
| ], | |
| "last": "Harkin", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "University of California", | |
| "location": { | |
| "settlement": "Los Angeles" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [ | |
| "Erwin" | |
| ], | |
| "last": "Reifler", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Mr", | |
| "middle": [], | |
| "last": "Victor", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "University of California", | |
| "location": { | |
| "settlement": "Los Angeles" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "H", | |
| "middle": [], | |
| "last": "Yngve", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "University of Washington", | |
| "location": { | |
| "settlement": "Seattle" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Yehoshua", | |
| "middle": [], | |
| "last": "Bar-Hillel", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "University of Chicago", | |
| "location": { | |
| "settlement": "Chicago" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Mr", | |
| "middle": [], | |
| "last": "Jay", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "University of Chicago", | |
| "location": { | |
| "settlement": "Chicago" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "W", | |
| "middle": [], | |
| "last": "Forrester", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "Research Associate, Research Laboratory of Electronics, Massachu-setts Institute of Technology", | |
| "institution": "", | |
| "location": { | |
| "settlement": "Cambridge" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "N", | |
| "middle": [], | |
| "last": "William", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Massachusetts Institute of Technology", | |
| "location": { | |
| "settlement": "Cambridge" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Locke", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Massachusetts Institute of Technology", | |
| "location": { | |
| "settlement": "Cambridge" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Vernon", | |
| "middle": [], | |
| "last": "Tate", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Jerome", | |
| "middle": [ | |
| "B" | |
| ], | |
| "last": "Wiesner", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Mr", | |
| "middle": [], | |
| "last": "Dudley", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "A", | |
| "middle": [], | |
| "last": "Buck", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge Mr. A. Craig Reynolds, Jr., Endicott Laboratories", | |
| "institution": "", | |
| "location": { | |
| "region": "I.B.M., Endicott, N. Y" | |
| } | |
| }, | |
| "email": "" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "", | |
| "pdf_parse": { | |
| "paper_id": "1952", | |
| "_pdf_hash": "", | |
| "abstract": [], | |
| "body_text": [ | |
| { | |
| "text": "THE FOLLOWING is a report on the proceedings of the first MT Conference, held at the Massachusetts Institute of Technology, Cambridge, Mass., June 17-20, 1952 , and my own reactions. 1 At the Conference individuals working on MT in this country and in England met for the first time and presented their different approaches. A detailed list of participants appears on the next page. The important point is that at this Conference linguists and electronic engineers joined for the first time to survey the linguistic and engineering problems presented by MT. At the end of the Conference it was the general impression of the participants that, for certain types of source material, a mechanization of the translation process is now a distinct possibility. Thus Dr. Warren Weaver's ideas about the possibility of MT in our time ceased to be a dream and moved into the realm of reality.", | |
| "cite_spans": [ | |
| { | |
| "start": 124, | |
| "end": 158, | |
| "text": "Cambridge, Mass., June 17-20, 1952", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 183, | |
| "end": 184, | |
| "text": "1", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "As a matter of fact, the engineers envisaged the creation of pilot machines within the next few years; that is, machines with limited storage for the translation of a limited quantity of scientific material from a foreign language into intelligible English, built for the purpose of convincing the general public and, especially, foundations and other organizations able to support new ventures, of the feasibility of MT, in order to obtain the funds necessary for further research and improvements.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The Conference was ably organized by Dr. Y. Bar-Hillel of the Research Laboratory of Electronics at M.I.T. Half a year earlier Dr. Bar-Hillel had visited the different groups working on MT in this country and published an excellent REPORT ON THE PRESENT STATE OF RESEARCH ON MECHANICAL TRANSLATION. 2 There can be no doubt that much of the success of the Conference was due to Dr. Bar-Hillel's efforts, and it is, I believe, no overstatement to say that MT, if and when it materializes, will 1 This report was written in July, 1952. Opinions and facts are of that date.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "2 AMERICAN DOCUMENTATION, 2:229 -237, 1951. be very much indebted to him.", | |
| "cite_spans": [ | |
| { | |
| "start": 2, | |
| "end": 43, | |
| "text": "AMERICAN DOCUMENTATION, 2:229 -237, 1951.", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The Conference decided that the papers of the participants should be published together with the discussions. 3", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Of greatest interest to the Conference was Dr. Booth's report on the translation experiments he and Dr. R, H. Richens had programmed on a computer in London. Dr. Warren Weaver had previously, in his first memorandum on MT (July 15, 1949) , referred to their work. According to him \"their interest was, at least at that time, confined to the problem of the mechanization of a dictionary which in a reasonably efficient way would handle all forms of all words.\" In a longer paper, SOME METHODS OF MECHANIZED TRANSLATION, which Dr. Booth submitted to the Conference he and Dr. Richens explain their approach. The translation they envisage is a word-for-word translation maintaining the word order of the input text and, in the case of multiple meanings, supplying alternative English equivalents. The machine determines by itself the stems and endings of the words of the input text and compares them with the entries in its separate stem and ending memories. These furnish not only the (often multiple) English equivalents for the input words, but also the (sometimes multiple) grammatical meanings involved. The latter are indicated in the output of the machine by abbreviations of the terms for the grammatical meaning concerned. At present only scientific material is considered for MT. Idioglossaries are used for the various fields, which means a considerable decrease in the number of possible meanings of each technical 3 Lack of sufficient funds has prevented the carrying out of this plan. However, a publisher has now been found for a volume of up-to-date essays reflecting present thinking on MT. This volume is scheduled to be published in the Prof. Stuart C. Dodd, Director, Washington Public Opinion Laboratory, University of Washington, Seattle term and an appreciable reduction both in the amount of storage required and in the access time. A number of sample products of this machine show the degree of intelligibility of the mechanical translation product and demonstrate how much this solution of MT leaves to the interpretation of a post-editor. There can be no doubt as to the value of Richens' and Booth's approach. It is, however, as they themselves are, I believe, very ready to admit, still far from the ideal of MT which I would define as follows: A complete mechanization of the translation process -that is, a mechanical system which, without the intervention of either a pre-or post-editor, outputs translations satisfactory with regard to both semantic accuracy and intelligibility. 4 Some of the participating linguists indicated in private conversations that the samples of automatic dictionary output were unintelligible to them. My own impression is that the time required for the interpretation of the meaning of the output of this machine will be a serious factor in the evaluation of its practicality. This time has to be added to the time required by the machine itself for its operations. People who know classical Chinese will, for obvious reasons, have less difficulty than others with the interpretation of the products of this machine.", | |
| "cite_spans": [ | |
| { | |
| "start": 222, | |
| "end": 237, | |
| "text": "(July 15, 1949)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 2511, | |
| "end": 2512, | |
| "text": "4", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Automatic Dictionary", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\"Word-by-Word\" or \"Block-by-Block\" Translations Other very valuable contributions were made by Professor Victor A. Oswald, Jr., of the UCLA who, together with Stuart L. Fletcher, Jr., had previously published PROPOSALS FOR THE MECHANICAL RESOLUTION OF GER-MAN SYNTAX PATTERNS. 5 In his conference paper WORD-BY-WORD TRANSLATIONS Dr. Oswald exemplified the inadequacies of such translation, even going so far as to assert that such a \"translation is literally impossible.\" He suggested instead \"block-by-block transverbalizatlon, in which process, problems of syntactic ambiguity are solved by the connection of syntactic segments with each other, and the fluid German word order is resolved into a rigid English sequence.\" This he had previously demonstrated in the PROPOSALS, \"...and,\" he added, \"we now know that a recognition of syn-4 See my chapter in the volume mentioned in footnote 3. LANGUAGE FORUM, 36:1 -24, 1951 tactic connection can be built into the 'memory' of machines of the high speed computer type.\"", | |
| "cite_spans": [ | |
| { | |
| "start": 277, | |
| "end": 278, | |
| "text": "5", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 892, | |
| "end": 922, | |
| "text": "LANGUAGE FORUM, 36:1 -24, 1951", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Automatic Dictionary", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Another important suggestion made in his paper and elaborated in a second paper entitled MICROSEMANTICS is his \"micro-glossariesglossaries which will reduce the range of choice of meaning from a bewildering multiplicity to a matter of -at the most -two or three.\" It has to be emphasized here that on every page of almost every scientific text scientific terms are rare islands in an ocean of general language. Consequently his scheme envisages \"microglossaries\" for the non-technical vocabulary of a whole domain of a particular science. This may reduce the number of non-grammatical meaning alternatives of the general language portions of scientific material in a number of cases. In the majority of cases, however, the non-grammatical incident meaning i.e., the particular meaning of the word in a given context, of these portions of the vocabulary is by no means determined or generally definable by the branch of science to which the material belongs, but has to be inferred from the meaning of co-occurrences of the narrow context. Therefore, although \"micro-glossaries\" (for which I suggested the obviously better term \"idioglossaries\" -it is also preferable to speak of \"idiosemantics\" rather than of \"micro-semantics\") will certainly play a significant role in the ultimate solution of MT, in the case of scientific source material we are still faced with all the problems of multiple non-grammatical meaning presented by general language. Microglossaries \"could,\" as Professor Oswald says, \"serve to replace a team of specialists (on the post-editor side) in our proposed process of MT.\" But they will, I am afraid, not enable us to dispense with a human editor or editors for general language problems, whether on the input or on the output side, or on both sides of the MT assembly line. Moreover, Professor Oswald is well aware that \"It is possible that it might be prohibitively expensive* to produce such glossaries.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Idio-Glossaries", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Of the greatest importance for the development of MT will be a conference paper by Professor William E. Bull of the UCLA, entitled PROBLEMS OF VOCABULARY FREQUENCY AND DISTRIBUTION. He exposes a number of \"fallacies which are current in most discussions of word frequencies\" From this highly techni-cal paper I quote only the following passages of great relevance for the problem of \"macro-\" and \"micro-glossaries\": \"There exists no scientific method of establishing a limited vocabulary which will translate any predictable percentage of the content (not the volume) of heterogeneous material. An all-purpose mechanical memory will have to contain something approaching the total available vocabulary of both the foreign (original) language and the target (final) language. In order to cover most semantic variations several millions of items would be needed. At the present time we have no machine which can manage such a number at a profitable speed.\"", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Vocabulary Frequencies and Distribution", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\"A micro-vocabulary appears feasible only if one is dealing with a micro-subject, a field in which the number of objective entities and the number of possible actions are extremely limited. The number of such fields is, probably, insignificant.\"", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Vocabulary Frequencies and Distribution", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\"The limitations of machine translation which we must face are, vocabularywise, the inadequacy of a closed and rigid system operating as the medium of translaition within an ever-expanding, open continuum.\"", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Vocabulary Frequencies and Distribution", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Operational Syntax and Teaching Foreign Languages", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Vocabulary Frequencies and Distribution", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Extremely valuable not only for MT, but also for all those interested in improving the teaching of languages is Professor Bull's second paper entitled TEACHING FOREIGN LANGU-AGES. I can here only quote some of the important suggestions made in his paper:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Vocabulary Frequencies and Distribution", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\"In teaching languages we should either replace rules by operational instructions or spell out in simple terms the operations necessary to make a rule work. I should like to stress in this connection, that the signs which may be used in teaching (and in the instruction of a machine) do not necessarily have to have any logical connection with the meaning. I shall give just two examples from Spanish. First, there are two verbs in Spanish commonly used to translate an English locative \"to be\": estar and haber. They are synonymous and even the educated native does not know what determines his choice. The signal is fundamentally non-semantic and the result of useless specialization in form usage. The problem, however, can be solved both for the machine and the student by isolating the fact that \"the\" in English takes estar and \"a\" takes haber.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Vocabulary Frequencies and Distribution", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The man is here. El hombre esta aqui.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Vocabulary Frequencies and Distribution", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "A man is here. Hay un hombre aqui.\" 6", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Vocabulary Frequencies and Distribution", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "It is Interesting here to note that Professor Bull's rule is perfectly applicable to the use of modern Chinese (haber). In the first case one cannot use , in the second case one has to use it. Incidentally, Dr. Bar-Hillel also strongly advocates the development of what he calls \"operational syntax\" for language teaching as well as for MT.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Vocabulary Frequencies and Distribution", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Other important statements in Bull's paper are the following:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Vocabulary Frequencies and Distribution", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\"The total volume of the high frequency words is established by counting their uses with the words included in the selection and all their uses with the rare words excluded from the selection. The student, consequently, who learns this vocabulary is over-supplied with cement and undersupplied with things to be cemented together. He is like a builder who is given ten tons of cement and 500 bricks and told to build a home. If he keeps his proportions proper he has to be contented with an elegant privy. I submit that this is one of the major sources of irritation and frustration in our elementary courses in foreign languages. The reason our students cannot say anything much after a year of language is not because they haven't studied; they haven't_got_a vocabulary whose proportions permit them to say anything but the obvious banalities.\" (The underscoring is mine.) \"The principle of excessive repetition cannot be sustained by the evidence of how a native is forced to learn his own language. This suggests strongly that we should increase the number of items given to the student and decrease, if possible, the number of repetitions of high frequency vocabulary.\"", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Vocabulary Frequencies and Distribution", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "In his conclusion Professor Bull suggests the following points for consideration in the improvement of language teaching: \" (l) the abandonment of outmoded elementalism, and research directed at language as a structural whole (2) a clear analysis of what is actually mechanical in language (3) the description of what the native's language-feel actually is (4) the substitution of operational instructions, whenever necessary for abstract rules (5) research to discover the mechanical signposts which are guides to usage (6) a new approach to the selection and teaching of vocabulary based on demonstrable facts\"", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Vocabulary Frequencies and Distribution", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Of the many valuable suggestions made by Professor Leon Dostert of Georgetown University I would especially like to mention one which will certainly become an important feature of future MT. Describing his experiences in multiple translations, he stressed the advantage of a \"pivot language\" or \"pivot languages.\" General MT (mechanical translation from one into many languages), he said, should be so developed that one translates first from the input language into one \"pivot\" language (which in our case will, most likely, be English) and from that pivot language into any one of the output languages desired. This will, I believe, be very beneficial for MT, as will become clear from the following.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Pivot Languages", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Professor Stuart C. Dodd of the University of Washington in Seattle addressed the Conference on MODEL TARGET LANGUAGES, (i.e., a regularized form of the languages into which one translates). His paper caused a very lively discussion as a result of which I can say that \"model TL-s,\" especially his \"model target English\" will constitute an important item in the mechanization of the translation process. As I pointed out in the first of my two papers (MT WITH A PRE-EDITOR AND WRITING FOB MT), if we aim at a practical solution of MT, then we can interfere neither with the language nor the conventional spelling (speaking here entirely with respect to alphabetized languages) of the original language. But on the output side we can, within certain definable li mits, plan the form of the output language. We can put a selected vocabulary and a regularized morphology and syntax into the machine and, moreover, within the limitations of intelligibility, adjust the final language to certain peculiarities of each of the original languages.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Model Target Languages", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Now in General MT, if we do not work with a \"pivot language,\" we shall (except in the case of original languages like Chinese and Japanese which by nature are very regular) in every case be faced with a mechanical correlation between one irregular and one regularized language. But if we do use a pivot language, then only at the first step will this be the case; that is, in the MT from a natural language into the pivot language. From here on, however, -that is, in the MT from the pivot language into any of the model output languages -we would in every case have a mechanical correlation between two regularized languages. Thus the use of a pivot language in General MT as suggested by Professor Dostert will mean a further simplification of the engineering problems involved.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Irregular Original Language -Model Pivot Language -Model Output Language", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "In my paper quoted above I also demonstrated how the graphic indication by a human agent of certain types of grammatical meaning in the input text might enable the machine to determine incident non-grammatical meaning. Drs. Bull and Oswald, however, in their papers foresaw the possibility that a machine might be designed to determine grammatical meaning by itself, on the basis of nothing more than the conventional graphic form of input texts. If this is possible, then that kind of pre-editorial work which my idea necessitates can be dispensed with. It will mean much for MT if it can be demonstrated that operational instructions can be abstracted from a language on which we can base the programming of a machine for the mechanical determination of certain types of grammatical meaning. But even so it is important to point out the following: a) even if this is possible for some types of grammatical information, it may not be possible for other types. In his MICROSEMANTICS Dr. Oswald mentions one kind of grammatical information for which he can -at least for the present -see only a human supplier. He says: \"The German system of noun compounding is such that a glossary based on the graphic forms would be both unwieldy and grossly inefficient because of unnecessary repetition. Almost any sequence of nouns in German not syntactically connected is automatically made into a compound, and your German noun strays gaily about appearing now as the \"head\" and now as the \"tail\" of a compound .... In a word, you must break up German compounds if you want to make any sort of efficient German-English glossary.... We know no mechanized process by which this could be accomplished, but an intelligent....pre-editor could indicate the dissection for any sort of context.\" 7 b) even though it is possible for some languages, it may not be possible for some others.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Mechanical Abstraction of Grammatical-Information", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "c) the machinery required may be so complex and expensive that we may ultimately prefer to have a human agent indicate the relevant grammatical information of the input text by some system of symbolization (pre-editor).", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Mechanical Abstraction of Grammatical-Information", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "d) if, as in the case of German compounds (see under a), no mechanized process can supply the information relative to one grammatical situation, so that this information has to be supplied anyway by a pre-editor, then the latter might as well add \"seam-signals\" to indicate the position of the \"seam\" (Oswald's \"fracturesurfaces\") in different types of compounds. The same signal would thus serve to indicate more than one type of grammatical meaning. This might result in a simplification of the mechanism designed for the determination of grammatical meaning because then the machine has more instructions on the basis of which to supply less information.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Mechanical Abstraction of Grammatical-Information", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "A most serious objection to my suggestion of a mechanical determination of incident nongrammatical meaning was voiced by Dr. Bar-7 Shortly after distributing my report on the conference I completely solved this problem of the mechanical dissection and identification of all predictable and unpredictable compounds. A detailed description of this solution, first reported in my SIMT Nos. 6 & 7 (mimeographed) will be included in the forthcoming volume mentioned in footnote 3.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Mechanical Determination of Incident Non-Grammatical Meaning and the Limited Storage Capacity of the Mechanical Memory", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Hillel. He said that such a plan would require a storage of billions or trillions of entriesobviously quite impossible to achieve. However, appearances are misleading here. Before I can show this, I have first to introduce a few new concepts:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Mechanical Determination of Incident Non-Grammatical Meaning and the Limited Storage Capacity of the Mechanical Memory", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "In the following I shall call \"clue-sets\" a set of co-occurrent words of which one or one group \"pinpoints\" the meaning of the remainder. I shall name \"pinpointers\" the pinpointing words and \"pinpointees\" those whose meaning is pinpointed by such \"pinpointers.\" Furthermore, I wish to remind the reader of the phenomenon of \"Shared Transferred Meanings\" discussed in # H/6 of my first paper on mechanical translation and of the vast possibilities of \"Pseudo-One-To-One Correlations\" exemplified in my second Conference paper. Lastly I shall speak about \"Pinpointees with a Manageable or Unmanageable Number of Pinpointers\" and about \"Pinpointee Meanings Stable or Unstable in the Light of Source-Target Semantics\" (I beg the indulgence of the reader for the freak terms \"pinpointer\" and \"pinpointee.\" I could not think of any other terms more \"to the point.\") Now Dr. Bar -Hillel's objection remains valid only if we are thinking of putting into the mechanized memory all possible clue-sets. This is, however, neither intended nor necessary. We have to consider here the following facts:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Mechanical Determination of Incident Non-Grammatical Meaning and the Limited Storage Capacity of the Mechanical Memory", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "1. Each set of two languages shares a considerable number of semantic parallels (shared transferred meanings). For example English will which, like Chinese , is used in the sense of \"to want, to wish\" and also as an auxiliary verb, expressing future; French \u00e7a va, German es geht and Chinese , meaning \"to go\" and also used in the sense of \"that does\" or \"that will do\"; Latin noli, \"don't,\" a contraction of non voli, meaning \"not want,\" and Chinese , meaning \"not want\" and \"don't\"; etc., etc.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Mechanical Determination of Incident Non-Grammatical Meaning and the Limited Storage Capacity of the Mechanical Memory", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "2. In an extremely large number of cases a literal translation, though resulting in an unaccustomed output form, is still perfectly intelligible either in the narrower or in the wider context. For example, in playing Chinese chess, a player may say ; which even in its literal translation, \"I eat your elephant\" (I take your elephant; the elephant is something like the bishop in Western chess), is perfectly intelligible to the English reader. We are in very many cases able to create artificial one-to-one correlations by selecting from the available output alternatives one which, though it may be customary or \"good\" only for certain context, is still intelligible in others. For example, Chinese , \"to create, make, do, act, etc.\", is also used in contexts where the English translator usually prefers to render it by forms of the verb \"to be.\" If we translate \"make\" also in these contexts, the result will often be horrible for the English hearer or reader, but it will still be intelligible. Thus \"he is a teacher, student, father, son, etc.,etc.\" would appear in the English translation as \"he make teacher, student, father, son, etc.\", which in its context, for example in answer to questions meaning something like \"what is his profession, position, what is he doing? etc.\" or when discussing somebody's duties in relation to his position, will be perfectly intelligible. A speaker of standard English does not need to learn pidgin English in order to understand what \"this master makee teacher\" (this gentleman is a teacher) means. 3. In every language there is a large number of words which may co-occur with a large number of other words \"pinpointing\" their incident meanings, but among these we have to distinguish several groups: a) \"Pinpointees\" whose meanings in the light of source-target semantics (semantic relationships between the pair of languages) are the same with all \"pinpointers,\" either in fact (semantic parallel, cf. point 1} or in terms of artificial one-to-one correlations (cf. point 2). Here no clue-set entries are necessary. The number of possible \"pinpointers\" is here, of course, of no consequence whatsoever for MT. For example German kaufen \"to buy\", verkaufen \"to sell\", schreiben \"to write\", essen \"to eat\", in terms of German-English and German-Chinese semantics. b) \"Pinpointees\" the number of whose \"pinpointers\" is comparatively small and whose meanings in the light of source-target semantics are, in terms of points 1 and 2 above, different with all \"pinpointers.\" Here all cluesets should and can be entered into the mechanized memory. c) 'Pinpointees\" the number of whose \"pinpointers\" is large and whose meanings in the light of source-target semantics are, in terms of points 1 and 2, the same in the case of a very large number of \"pinpointers,\" but different in the case of a small number of \"pinpointers.\" Here no clue-set entry is necessary in the first case, whereas in the second all clue-sets should and can be entered. d) \"Pinpointees\" the number of whose \"pinpointers\" is large and whose meanings in the light of source-target semantics are, in terms of points 1 and 2, the same in the case of a comparatively small number of \"pinpointers,\" but different with regard to a large number of \"pin-pointers.\" Here no clue-set entry is necessary in the first case, whereas for the second the decision has to be deferred until we know more about the size of the total residual problem. e) \"Pinpointees\" the number of whose \"pinpointers\" is large and whose meanings in the light of source-target semantics are, in terms of points 1 and 2, different with regard to different groups of \"pinpointers.\" Here we can certainly enter all clue-sets relative to one of the groups, preferably the group with the largest still manageable number of \"pinpointers,\" whereas for the remainder the decision has to be deferred until we know more about the size of the total residual problem. f) \"Pinpointees\" the number of whose \"pinpointers\" is large and whose meanings in the light of source-target semantics are, in terms of points 1 and 2, different with regard to every \"pinpointer\" (this situation will be either rare or not occur at all). Here the decision has to be deferred until we know more about the size of the total residual problem.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Mechanical Determination of Incident Non-Grammatical Meaning and the Limited Storage Capacity of the Mechanical Memory", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Thus wherever transferred meanings are shared or wherever we can artificially create one-to-one correlations, no consideration of \"pinpointers\" is necessary and, consequently, we need not worry about the entry of clue-sets. Wherever transferred meanings are not shared, or wherever we can not artificially create oneto-one correlations, and where the number of \"pinpointers\" is comparatively small, we certainly can enter all clue-sets. Thus we are ultimately concerned only with the residual problem of those cases where \"pinpointers\" have to be considered and are very numerous. No research has ever been done for any set of two languages to determine the size of the residual problem. It is, therefore, not possible to decide on its treatment at present. If it still required more than, say, 10 million entries, one would naturally hesitate to consider recording in the mechanized memory. What is important, however, is that, assuming the residual problem required too many entries to permit mechanization, the machine would leave only this residual group of multiple meanings to a preor post-editor. The editor would have much less editing to do and in the case of a posteditor the difficulty of semantic determination might well be diminished to a degree he would certainly appreciate: the larger the number of semantic decisions the machine makes for him, the clearer the output context he has to consider for the solution of the remaining riddles! Certainly, in MT wherever mechanization is practical, it should be carried out!", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Mechanical Determination of Incident Non-Grammatical Meaning and the Limited Storage Capacity of the Mechanical Memory", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "In this context I should like to add some remarks to the problem \"pre-editor versus posteditor.\" In my first two papers on MT 1 burdened the pre-editor not only with the signalization of the grammatical, but also with that of the incident non-grammatical meaning; that is, wherever source-target semantics presented a problem of multiple meaning. In #81 of the first paper I had actually previously considered the alternative possibility of using a post-editor to whom, in the case of multiple meanings, the machine would supply the various alternatives from which he would have to make the correct selection. I had said there that from the point of view of complete mechanization this may seem to be preferable because then no human factor would interrupt the purely mechanical side of MT. However, from the point of view of MT as a whole, using a pre-editor is still much quicker for the following reasons: whereas the reader of the original text (i.e., pre-editor) has to select the meaning that \"makes sense\" in an original context which is completely intelligible to him, the output text reader (i.e., post-editor) has to do this in an output context which will necessarily contain a large number of nondistinctive words with transferred meanings different from those of the corresponding original language words, that is in_a context_that will often not be clear.\"", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Pre-editor Versus Post-editor", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Dr. Bar-Hillel, on the other hand, advocates the determination of such incident meanings by a post-editor and has found much support for his idea. As a matter of fact, at this early stage of MT research I, too, cannot completely rule out the possibility that a MT post-editor (not to be confounded with a general post-editor concerned with stylistic improvements of the output text) may be necessary for the solution of at least some of the semantic problems involved.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Pre-editor Versus Post-editor", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Professor Oswald in his WORD-BY-WORD TRANSLATION voiced his scepticism concerning both the pre-and the post-editorial approach. \"I do not believe,\" he says, \"that his (i.e., Reifler's) combination of pre-editor with a mechanical dictionary constitutes the ultimate solution of our problem. In fact, I am of the opinion that we must grapple with the problem precisely at the point where Mr. Reifler abandons it. His proposals are most enlightening for the solution of problems of general language, but he has excluded problems of specific language (the jargon of medicine, mathematics, linguistics, geology, etc.) from the domain of mechanical solution. We shall be much closer to the realization of mechanical translation if we can mechanize the components of his \"mechanized'' dictionary....A pre-editor can do much to simplify syntactic connection for mechanical 'digestion,' but I do not see how, as an operator in the FL (i.e., foreign or original language), he can effectively guide either the machine, or the machine plus a post-editor, through the mazes of multiple meaning on the TL (target or final language). Nor do I think we can hope for much accurate help from one monolingual post-editor or even from one bilingual consultant. What has been overlooked is the fact that the competence required in the post-editor, even if he be bilingual, is only partially linguistic. The real prerequisite for him is an intimate knowledge of the field to which the translated text pertains\" (pp. 3-5).", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Pre-editor Versus Post-editor", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Apart from the fact that I have in no way \"excluded problems of specific language...from the domain of mechanical solution\" (I am fully aware of the urgency of the translation of scientific material, but would point out that even in such material we have to solve problems of general language), I fully agree with Professor Oswald. But he had, when he wrote his paper, not yet seen my third paper (the first submitted to the Conference) in which I indicated my radical departure from my previous position, demonstrated the possibility of mechanizing the determination of incident non-grammatical meaning on the basis of information relative to certain types of grammatical meaning, and limited the work of the pre-editor to the signalization of these types of grammatical meaning. Both Drs. Oswald and Bull have, on the other hand, mentioned the possibility that the determination of incident grammatical meaning may be mechanized. If this can be done, then there would remain only the question whether the solution of all multiple meaning problems (in case no portion of this problem can be mechanized) or of the semantic problems left over by the machine is -from the point of view of all-round practicality -better done by a preor a post-editor. I still feel that this task is easier for the pre-editor. The post-editor is faced with a non-conventional form of output context in which he has to make a selection from each of a number of conglomerations of output alternatives in consideration of one or more other conglomerations of output alternatives. He does, in fact, not fully understand the narrow output context before he has made at least some correct selections. The preeditor, on the other hand, is confronted with a familiar linguistic medium without any conglomerations of alternative words and under-stands the contexts before he is informed about the existence of a multiple meaning problem in terms of source-target semantics and before he has chosen the appropriate supplementary signal from the dictionary entry supplied by the mechanized dictionary. If we assume that a large portion of the multiple meaning problems can be solved mechanically along the lines 1 have suggested and that the pre-editor would thus be faced only with the residual semantic problems, then the combined man-machine procedure would be something like the following. The pre-editor sends the original text into the dictionary mechanism. In all cases of multiple meanings in which the dictionary mechanism can itself determine the incident meaning and supply the appropriate output equivalent on the basis of the supplementary grammatical signals which the pre-editor has added to the conventional graphic form of the original text (or on the basis of the grammatical information Bull's and Oswald's \"grammar mechanism\" has abstracted and supplied to the dictionary mechanism), the pre-editor would never have to know that multiple meanings in terms of sourcetarget semantics are involved. The machine would do the work without giving any hint that there are such multiple meaning problems. In the case of a residual problem, however, the machine would in every case notify the preeditor in some way and supply him with a dictionary entry (in his own language!) indicating the meaning alternatives in the light of sourcetarget semantics. From these the pre-editor would have to choose and then add the appropriate supplementary signal to the portion of the input text involved. As pointed out above, he can make such a choice much quicker than a post-editor because he is dealing with a familiar linguistic medium and understands the output context before he makes his choice.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Pre-editor Versus Post-editor", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "I should like to add that I am keeping an open mind with regard to this problem of pre-editor versus post-editor. It is, in fact, quite possible that, in terms of the time and money spent on linguistic and engineering research (linguistic research is probably less expensive than engineering research), mechanical complexity and construction time, speed and accuracy of translation, etc., etc., the optimum may be reached in an arrangement in which a pre-editor signalizes certain types of grammatical information, the machine abstracts some other types of grammatical information and on the basis of this information from two sources determines certain types of incident non-grammatical meaning and reshuffles the word order. A posteditor then solves the residual semantic pro-blems on the basis of an output context which, because it does not contain too many clusters of alternatives, is much clearer.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Pre-editor Versus Post-editor", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Professor Dostert suggested the early creation of a pilot machine or of pilot machines proving to the world not only the possibility, but also the practicality of MT. Since the time necessary for the creation of such machines is an important factor, it will be best to develop a plan based on the simplest possible conditions. When this problem was raised at the Conference, the general opinion seemed to be that the simplest conditions are found in the mechanical correlation of certain European languages (Germani) with the English language. I pointed out, however, that contrary to appearances, a German-into-English scheme can not in the least compete with a Chinese (or Japanese) into English scheme. In the case of these two languages nature has already provided us with highly regular languages. Moreover, both in morphology and syntax Chinese and English happen to have more in common than German (or any other European language) and English. If we put into the translation mechanism a regularized Englich which is, furthermore, within the limitations of intelligibility, adjusted to certain peculiarities of Chinese, we have an ideal situation: a correlation between two regular and in many respects very similar languages. It is true that -as was stressed at the Conference -certain government agencies may be readier to supply the funds necessary for further research and improvements if the first pilot machine is designed for mechanical translation from Russian into English. But such a machine will be more complex and more expensive and the work necessary for its creation more time-consuming than in the case of a Chinese-English MT unit.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Pilot Machines", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Thus the first pilot machine should, I feel, be programmed for a MT from Chinese into Englich. Moreover, if we want to go further and show the possibility and practicality of General MT (mechanical translation from one into many languages) on the basis of the concept of \"pivot languages\" as suggested by Dr. Dostert, our simplest proposition would be one in which we add to the Chinese-English unit a second unit for the translation of the English output of the first unit into Japanese. Then we would have a mechanical correlation merely between a regularized language (English) and another language (Japanese) which by nature is highly regular.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Pilot Machines", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The Conference ended on an optimistic note with the suggestion by Professor Booth that the next conference be held in London.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Pilot Machines", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Chinese Characters Versus Alphabetization I should like to add here a valuable suggestion which has come to me from Dr. Fang-kuei Li. With regard to languages with a non-alphabetic script I had hitherto thought of making use of an alphabetized form. I had pointed to the fact that, wherever different alphabetization systems have been suggested or are actually used, the graphio-semantically most distinctive one would be most beneficial for MT. For Chinese this would be the I.R. (Interdialect Romanization). But even in this romanization some additional differentiation is necessary in order to further reduce the still large number of homographs. Dr. Li suggested that, since even the I.R. requires further adjustments for purposes of graphio-semantic distinctiveness, it may be worthwhile to consider the development of sino-foreign MT on the basis of the Chinese characters themselves, which are graphiosemantically more distinctive than the I.R. He added that he had heard that a machine supplying the corresponding characters for the Chinese telegraph code numbers has already been developed in this country. There should be no reason why a machine which reverses this process could not be built. A pre-editor could add the supplementary grammatical signals just as well to a Chinese character text as to an alphabetized form of this text. The supplementary signals would be typed into the character-(code) number machine together with the characters to which they refer. Such an approach would eliminate the transcription into an alphabetization and thus save time. 8", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Pilot Machines", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "TEACHING FOREIGN LANGUAGES, p.3. For the second example, see the original.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "For dates and references to Dr. Reifler's papers on MT, see Vol. I, No. 1 of MT, March 1954.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| } | |
| ], | |
| "back_matter": [], | |
| "bib_entries": {}, | |
| "ref_entries": { | |
| "FIGREF0": { | |
| "num": null, | |
| "uris": null, | |
| "type_str": "figure", | |
| "text": "fall of 1954 jointly by the Technology Press of M.I.T, and John Wiley & Sons. It is being edited by A. D. Booth and W. N. Locke. Participants in the Conference on Mechanical Translation Dr. A. D. Booth, Director, Electronic Computer Section, Birkbeck College, London Prof. William E. Bull, Department of Spanish, University of California, Los Angeles" | |
| } | |
| } | |
| } | |
| } |