Báo cáo khoa học: "Some New Terminology"

Chia sẻ: Nghetay_1 Nghetay_1 | Ngày: | Loại File: PDF | Số trang:0

Thêm vào BST

Báo xấu

51
lượt xem 1
download

Download Vui lòng tải xuống để xem tài liệu đầy đủ

MT research requires cooperation between engineers and linguists. It is important, therefore, to develop a uniform linguistic terminology that can be understood and used by engineers. Furthermore, it is necessary that linguists develop an understanding of the engineering problems involved.

Chủ đề:

Bình luận(0) Đăng nhập để gửi bình luận!

Lưu

Nội dung Text: Báo cáo khoa học: "Some New Terminology"

[Mechanical Translation, vol.4, no.3, December 1957; pp. 52-53] Some New Terminology Erwin Reifler, University of Washington, Seattle, Washington M T research requires cooperation between engineers and linguists. It is impor- tant, therefore, to develop a uniform linguistic terminology that can be understood and used by engineers. Furthermore, it is necessary that linguists develop an un- derstanding of the engineering problems involved. The results of cooperation be- tween linguists and engineers working with the MT Pilot Model at the University of Washington are presented here. THE LINGUIST interested in pioneering in MT 3. Input Symbols include all contextual sym- h as to struggle with two difficult problems b ols that may appear in a source t ext. from the very outset: 1) the formulation of an 4. Output Symbols include: adequate linguistic terminology that can be un- a) Letter symbols of the target alphabet derstood and used by the engineer, and 2) an b) Symbols for the numerals understanding of the engineering problems in- c) Punctuation symbols v olved. During our eight years of MT re- d) Editing symbols — target symbols in- search at the University of Washington we have tended to aid in the interpretation of the MT h ad the great advantage of close cooperation product. Examples are subscript numbers between linguists and engineers. I wish to sub- which are attached to some target equivalents m it for discussion under the heading of "Ter- to pinpoint the field or fields of science to which minology" some of the results of this coopera- the scientific meanings of certain semantic units tion. of the source language belong. (The term "se- R ecent developments in MT research at the mantic unit" will be explained below.) University of Washington have necessitated the redefinition of some old linguistic terms and 5. Free Symbol — a contextual symbol pre- the formulation of some new ones. They con- ceded and followed by space. It is always cern the concepts of MT symbols, i.e., all meaningful and always used to symbolize both graphic symbols used in the machine translation grammatical and non-grammatical meaning. process. These MT symbols consist of the A n example is English 'I'. Control Symbols and Contextual Symbols. 6. Bound Symbol — a contextual symbol either 1. Control Symbols — MT symbols which, n ot preceded or not followed, or neither pre- coded into the machine memory, control cer- ceded nor followed by space. We distinguish t ain steps in the translation process. Since a) Left-bound symbols they are not contextual symbols, they appear b) Right-bound symbols neither in the input nor in the output. c) Twice-bound symbols 2. Contextual Symbols — the minimal contex- 7. Meaningful Bound Symbol — a contextual t ual constituents used to produce a material symbol used to symbolize: s timulus for a machine-operational step rele- a) Grammatical meaning, i.e., left-bound vant for MT, such as an alphabetic letter, a " s" in "father's, fathers", the right-bound " ' " numerical figure, a dollar sign, a punctuation in " 's" which indicates that the following "s" is mark, a single space. Contextual symbols a substantive ending, the twice-bound "o" in consist of Input Symbols and Output Symbols. " arterio-sclerosis."
N ew Terminology 53 b) N on-grammatical meaning, i.e.., the 12. Meaningful Bound Symbol Sequence — a l eft-bound "g" which distinguishes the meaning b ound symbol sequence used to symbolize: o f "pang" from that of "pan", the right-bound a) G rammatical meaning, i . e . , left-bound " s" which distinguishes the meaning of "span" " ren" in "children", and right-bound "be" in f rom that of "pan", the twice-bound "a" distin- " befall" which changes the intransitive meaning g uishing the meaning of "seat" from that of o f "to fall" into a transitive meaning, twice- " set." b ound ы в d istinguishing the grammatical mean- i ng of о писывать ' to describe' (imperfective c) B oth grammatical and non-grammatical a spect) from that of о писать ' to describe’ (per- m eaning, i . e . , right-bound " о " distinguishing f ective aspect). t he grammatical and non-grammatical meaning b) N on-grammatical meaning, i.e., left- o f о писать ' describe’ (perfective aspect) from b ound "et" distinguishing the meaning of "ballet" t hat of п исать ' write' (imperfective aspect), f rom that of "ball", right-bound "bl" distinguish l eft-bound “ я ” d istinguishing the grammatical i ng the meaning of "bleat" from that of "eat", a nd non-grammatical meaning of л омя ' break- t wice-bound "ur" distinguishing the meaning of i ng' from that of л ом ' crowbar', twice-bound " gourd" from that of "god". " ж " distinguishing the grammatical and non- c) B oth grammatical and non-grammatical g rammatical meaning of м ежду ' between' from m eaning, i . e . , left-bound "shore" in "sea- that of меду ' of the honey'. s hore", right-bound "sea" in "seashore", and t wice-bound "en" in "disentomb". 8. Meaningless Bound Symbol — a bound s ymbol not intended by the author of a source 13. Meaningless Bound Symbol Sequence — a t ext to symbolize anything, but treated as a b ound sequence not intended by the author of a s eparate entry by the MT planners in order to s ource text to symbolize anything, but treated o vercome engineering difficulties due to certain a s an individual entry by the MT planners in l imitations of the MT equipment. An English o rder to overcome engineering difficulties due e xample is the arbitrary left-bound final sym- t o certain limitations of the MT equipment. An b ol "n" in "misinterpretation" which consists E nglish example is the meaningless left-bound o f 17 letters. If, for example, the input equip- s ymbol sequence "ss" in "irreconcilableness" m ent cannot handle free symbol sequences w hich consists of 18 letters. The MT planners l onger than 16 letters, then "misinterpretation" w ould have to split this free symbol sequence m ay be split arbitarily into two constituents, i nto two arbitrary constituents containing 16 t he first of which contains the first 16 letters a nd 2 letters respectively, and enter them as w hile the second consists of only one letter. s eparate entries into the machine memory if T hese two constituents would then form two t he available input equipment cannot handle s eparate entries in the machine memory. f ree symbol sequences longer than 16 letters. 14. G roup o f F ree S ymbol S equences — a 9). Symbol Sequence — a sequence of contex- c omplete text or any part of a text, chapter, t ual symbols not interrupted by space. s ection, sentence or clause consisting of two o r more free symbol sequences which symbol- 10. F ree S ymbol S equence — a s ymbol se- i ze a meaning intended by the author of the q uence preceded and followed by space. A free s ource text. s ymbol sequence is always meaningful and is a lways used to symbolize both grammatical 15. A Semantic Unit — a single free or bound m eaningful s ymbol or symbol sequence, and a nd non-grammatical meaning. a ny group of free s ymbol sequences which is i diomatic in terms of source-target semantics. 11. B ound S ymbol S equence — a s ymbol se- q uence either not preceded, or not followed, or W ith the growth of MT development and the n either preceded nor followed, by space. We i ncrease in the number of MT pioneers it is d istinguish: b ecoming more and more important to achieve s ome uniformity in linguistic terminology for a) L eft-bound symbol sequence M T. I submit the above definitions for criti- b) R ight-bound symbol sequence c ism and suggestions. c) T wice-bound symbol sequence