Total Pageviews

Tuesday, 22 November 2011

CAN COMPUTER SPEAK? CAN COMPUTER SPEAK? LANGUAGE AND ARTIFICIAL INTELLIGENCE








Debaprasad Bandyopadhyay *

^ দেবপ্রসাদ বন্দ্যোপাধ্যায়


ABSTRACT


Speaking/Hearing Subject(S/HS) is  desiring  to be a machine:

Machine= f (S/HS)

 Non-algorithmic zones of Speaking/Hearing Subject with n-nary options are annihilated. Meaning is solidified, condensed. Corpse of corpus is clerically explored. Binary machine does not understand what “understanding” is. Do we not consider Russell’s paradox or Godel’s theorem here or are making fun with non-biodegradable and disposable ostentatious talking toys?
 The status of COMPUTATIONAL LINGUISTICS is questioned here from four perspectives: from the standpoints of (a) Philosophy of science, (b) Natural Science (mismatch between human Cognitive domain and machine algorithms) (c) Social Science (Plurilingual Planning), (d) Algocentricism (in contrast with post-formal subjective and substantive task of Linguistics). These problems are summarized as follows:


 A. PHILOSOPHY OF SCIENCE PROBLEM
THERE IS NOTHING CALLED “PEN-PAPER-CARD LINGUISTICS”
IF COMPUTER MANIPULATES LINGUISTIC DATA THROUGH THE “PEN-PAPER-CARD” METHOD
IS IT JUSTIFIED TO LABEL IT AS A SEPARATE DISCIPLINE?
WHAT IS THE EPISTEMOLOGICAL STATUS OF CORPUS LINGUISTICS???

B. NATURAL SCIENCE PROBLEM

MATCHING CONDITION BETWEEN “HUMAN COGNITIVE DOMAIN”
&
“MACHINE ALGORITHM”
[Identity in Difference between computer and human being]
[a] RUSSELL’S PARADOX
GODEL’S/ CHURCH’S THEOREM
[b] PROBLEM RAISED BY PENROSE (1990,1994)
&
SEARLE’S CHINESE ROOM PUZZLE
[C] COMPUTER’S HALTING PROBLEM
[d] FUZZINESS OF NATURAL LANGUAGE
[e] POST- FORMAL APPROACH THAT DENIES ANALYTICAL PROCEDURES PROPOSED BY STRUCTURALISTS
C. PROBLEM OF ALGOCENTRICISM
Metamathematical algorithmic procedural rules ignore the non-algorithmic constitutive rules (?) of human cognition. 

D. SOCIAL SCIENCE PROBLEM
PLURILINGUAL LANGUAGE PLANNING
ELECTRONIC CAPITALIST INTERVENTION IN LANGUAGES
THE EMPIRE OF HYPER-REAL, Simulation……..


 NEOLOGISMSAlgocentric Discourse, Electronic Capitalism
TWO INSTANCES OF NON-COMPUTABILITY

(1) FUZZY NUMERALS
One thing is certain that meaning is uncertain. What a paradox! Aporia! Invented numerals are fuzzy—indeterminacy prevails. An ostentatious toy was made—a fuzzy game.  (S)talker tried to depict some non-deterministic and uncertain phenomena revealed in the expressions of numerals by Bangla speakers. The points of arguments are as follows:
1. Human quantification is partly or fully different from analyst’s rigorous quantification.
2. The following corpus from Bangla has a peculiar nature of non-componentiality or they are rather prototypical.

This prototypical nature of fuzzy numerals cannot be handled in computational framework or even  in the Logical Form. The first set of data deals with idiomatic expressions like:
1. Sat-paMc  Seven-five “pros and cons”;
2. nOy-chOy nine-six “topsy-turvy”;
3. unis-biS, nineteen-twenty “trifle difference” jaHa bahanno taHay tippanno.
5. jaHa baHanno, taHai tippanno “Whatever is fifty-two, that is fifty-three” (i.e. 52=53) “A trifle difference does not count.”

The second problem may be termed as “one is not equal to one” problem. A rule of “one deletion” was proposed by Probal Dasgupta (1987).
5 dokanduTo shop-two-classifier.
6.dokanTa shop-singular-classifier.
Obviously, “one” is deleted in 6. However, Dasgupta mentioned that ‘one deletion’ is not true in the cases like
7. jOlTa, Water-classifier or
8. telTa oil-classifier. In Bangla, one cannot say
9. *EkTa jOl ‘One-classifier Oil”
However, there are some pragmatic cases where such expressions like 9 is possible. The Speaking  subject’s perception may still be “one” in those cases– it is ‘one’ as a mass body. Of course, this is not a deterministic physical ‘one’, but one as a whole. When any Bangla speaker says,
10. phrij theke jOlTa ano. fridge from water classifier
bring “Bring water from the fridge.”,
his/her intention is to refer “one bottle of water”. Therefore, ‘one” is there in the D-structure, but it is a fuzzy one. The concept of this fuzzy “one” may  be further illustrated in the following movement- transformations, where deterministic numeral expressions are changed to non-deterministic Determiner Phrases:
11.a) paMcTa five-classifier (definite)
11.b) goTa paMcek classifier( indef. ) five-one “more or less five “
12.a) paMcjon five-classifier (definite)
12.b) jona paMcek classifier( indef. ) five-one “more or less five “
13.a) paMckhana five-classifier
13.b) khan paMcek ,classifier( indef. ) five-one “more or less five “
14.a) paMc Hajar “five thousand”
14.b) Hajar paMcek thousand five-one. “more or less five thousand”
15.a) paMc lakh, five lacs
15.b) lakh paMcek lacs five-one “more or less five lacs”

Examples like 11-15 show those deterministic expressions in (a) and non-deterministic expressions in (b). Compared to (a), examples in (b) show the fronting of classifiers with subsequent morphophonemic change and an addition of /ek/ “one” to the specific numeral x. This one is not deterministic +-1, but this “one” has a range more or less than +- 1.
These Bangla numeral expressions show the world-views of the community concerned with a special reference to their psychophysical way of looking at things (perception) and ways of making order of things (understanding). Therefore, it is a hermeneutic problem that involves the relative gap between human perception and understanding in relation to their habitat. The range of +-1 is different in different persons belonging to different socio-economic classes or even it may be different in a single person in different psychosocial context. A game had been developed by his engineering students

(2) PERCEPTUAL AND GRAMMATICAL TIME
Perceptual time is supplemented by grammatical tense. Do we need to deploy tensor calculus here? That is ridiculous. Let us watch the watch in the time of eco-enemy technocratic society guided by technical rationality.
There are different types of time, (a) technological time as it is followed by watching the watch by selecting a norm (GMT); (b) grammatical time as it is revealed in the textbooks of Grammar; (c) perceptual time as it is perceived by the speaking subjects of a given speech community; (d) Scientific time as it is calculated by the locus of the perceived by deploying Tensor Calculus. The author showed that the prescriptions of grammatical time do not match with the perceptual time of different communities. The author argued that the tenses and aspects as categorized in the Bangla prescriptive grammars ignore the perceptions of the construct “time” by the speakers of the imagined community “Bangla” as the conjugational paradigm in the Bangla grammar book is the mimicry of English prescriptive grammatical paradigms, e.g.,
1.      ami tomar baRi kal jacchi
I your house tomorrow go present continuous
“ I am going to your house tomorrow”
2.      ami tomake kal rastay dekhechi.
I you dative yesterday road locative see present perfect
“I have seen you on the road yesterday.”
In the case of 1, all the Bengali-speaking informants perceived /jacchi/ as future tense (I will go..) and in the case of 2, /dekhechi/ is perceived as past tense (I saw) instead of prescribed “present perfect” (which is traditionally used to denote “immediate past), though prescriptive grammar categorically put an imperative for not using “present perfect” in the sense of past (cf. Chattopadhyay, 1939: 322). It is matter of wonder that there is no linguistic work so far on the perceptual time in Bangla.
This paper also concentrated on the semiotics of time in Cinema or particularly on the concept of time in the hyper-real, real (as it is symbolized) and in the ir-real. The author compared the difference between “reel” time and “real” time at the moment of perceiving hyper-real. As the author has dealt with the subjective perceptions of time in different locus and contexts by different individuals of different imagined communities, it is concluded that there is no generalized objective parameters for the historical a priori, “time”.
ANOTHER SIDE OF THIS NARRATIVE

Algorithms are used to simulate intonation: MUSICKING AND SPEAKING: DIFFERaNCE (SEGMENTING THE SUPRASEGMENTALS)   Download (.pdf)

APPENDIX
In the beginning of my career, I dare to ask Chomsky all about this man-machine equation; about impossibility of Artificial Intelligence; about his algocentric discourse that follows machine algorithm(especially the use of the term ‘computation’); about his model-theoretic approaches….

His answers are as follows:


For detailed discussion, kindly follow hyperlinks (blue-colored titles)  

·         2006. “Linguistic Cybercolonization”.  Globalization, Language, Culture and media. Eds. B.N.Patnaik, S.I.Hasnain, Simla Institute of Advanced Studies.(pp.146-187) ISBN81-7986-061-2  Download (.pdf)

·         2004. কম্পিউটার কি কথা বলবে? [Can Computer Speak?]”Prasanta Chattopadhyay ed. Kalodhvoni. XII:2-3 (pp.28-36)  Download (.pdf)


·         2001. পাড়া থেকে বিশ্ব বা বিশ্ব থেকে পাড়া : সাইবার কলোনির অভ্যুত্থান.” biSSaYon, gOnotOntro o tritio biSSo.(Proceedings of the Seminar on “Globalization”). Vol.II. Indian Academy of Social Science. Jadavpur University. (pp.3-8)  Download (.pdf)

·         2003. নানান সময়: চলচ্চিত্রেরও…” [Language, Time and Cinema] Silhouette. Vol. III. (pp.112-25).  Download (.pdf)

·         2000. Cybermultilingualism.Indian Journal of Applied Linguistics.  XXVI:2 (pp. 117-130). Delhi. ISSN  0379-0037.  Download (.pdf)

·         2000. শেষ মহাআখ্যান [Last Meta-Narrative]. proticcer SaHittotOtto. Kolkata :Alochona Chakra (pp.91-151).  Download (.pdf)

·         1999. “Bangla Numerals and Problems of Computability”.  2nd. International Conference on South Asian Languages, Punjabi University, Patiala. 8-10 January, 1999.http://linguistlist.org/pubs/papers/browse-papers-action.cfm?PaperID=7802

·         1999. Localization in Globalization Frontier. Vol.31. No.40.May 2-8, 1999 (pp.12-15)  RN 16516/68.  Download (.pdf)

·         1999. Electronic Capitalism.” Frontier. Vol. 31, No. 25. ( pp. 4-6).  Download (.pdf)

·         1997. “Scholars Differ : D.B. & T.M.M. On Probal Dasgupta & L. Khuchandani”. DLA News. 21:11 November ’97 (pg.3) RN KL/Tv(n)/116 ( A Discussion On the Review of  Khubchandani’s 1997. Revisualizing Boundaries; Introduction: Probal Dasgupta)

·         1997. “On Computational and Chomskyan Linguistic Theory”. Bahri, U.S. ed. Indian Journal of Applied Linguistics. Vol .XXIII, No.2 (pp.29-42). Jul-Dec’97 Delhi. ISSN  0379-0037.  Download (.pdf) 


DigiArt: Akhar Bandyopadhyay





No comments:

Post a Comment

Note: only a member of this blog may post a comment.