Fase 3. Formulación del modelo y conclusiones
5. CAPÍTULO V – ANÁLISIS DE LA INFORMACIÓN
5.1 ANÁLISIS PLAN DE ORDENAMIENTO TERRITORIAL DEL MUNICIPIO DE CHÍA
the late 1960s and the 1970s were the heyday of cognitivism. the cognitive approach permeated most branches of psychology. and apart from a vague feeling that in practice, if not necessarily in principle, cognitivism tended to ignore individual differ- ences, all seemed well. Indeed, the cognitive science movement of the 1970s, primed partly by money from the alfred P. sloan foundation in the united states, tried to make a more systematic connection between cognitive psychology, the key disciplines that had influenced its formation – aI and linguistics – and other cognate disciplines such as anthropology, education, neuroscience, and philosophy. It is not really clear that this movement forged relationships that did not already exist. However, one piece of work, on the edges of the cognitive science movement, signalled a new way of thinking about the relation between psychological, neurological and computational approaches to behaviour – david marr’s study of vision and particularly low-level vision.
marr originally studied neurophysiology and neuroanatomy, and developed an important and still current theory of cerebellar functioning in the late 1960s. He became interested in aI and moved to the aI laboratory at the massachusetts Institute of technology in the early 1970s. However, he was critical of the semantic information processing/blocksworld approach to computer vision that was current at mIt in the late 1960s and early 1970s. apart from problems that he identified with the details of the analyses developed, marr felt that, at least if it were to be regarded as providing a theory of human psychological function, aI research suffered from a confounding of science and engineering (the production of working computer programs). marr proposed that psychological functioning should be analysed at three levels, which he called computational, algorithmic and implementational.
the computational level is somewhat misleadingly labelled. It refers to an abstract analysis of the task or tasks that a psychological system performs. this is the level of analysis that marr found missing, or largely missing, in the standard aI approach to vision (and in much psychological work, too). the algorithmic level is, more or less, the standard abstract functional level of analysis of cognitivism, and the imple- mentational level relates the algorithmic analysis to specific neuronal functioning. In much of marr’s work, the algorithmic analysis uses neural network type computa- tions, similar to those found in connectionist models. Particularly at the lower levels of vision, this type of analysis is informed by what is known about the underlying neuronal architecture (for example in the retina).
marr’s approach to the relation between neuroanatomy and neurophysiology reflects standard thinking about, for example, the relation between sciences such as physics and chemistry. chemical laws and principles are discovered and supported in their own terms (by chemists doing chemistry). but it is also necessary to show, at least in principle, how they relate to the more fundamental physical mechanisms and processes that underlie them.
In his study of higher level visual processes (object recognition) marr’s work became both more speculative and less tied to theories of implementation. In marr’s approach, identifying gross brain areas that support certain functions is relatively uninformative. real understanding comes when a detailed algorithmic analysis can be related to the details of the underlying hardware. so although marr is described (for example in his Wikipedia entry) as the founder of cognitive neuroscience, it is not clear that the lessons of his work have properly been absorbed in any of the core disciplines that his research relates to (psychology, neuroscience, and aI), nor that it fully informs current research in cognitive neuroscience.
the cognitive science movement failed to entrench cognitivism’s position as the dominant movement in the study of psychological functioning. Indeed, cogni- tivism has suffered a number of setbacks since the early 1980s. the connectionist approach, which as has been mentioned was related to some of the algorithmic analyses developed by marr, became influential following the publication in 1986 of mcclelland and rumelhart’s two volume edited book on Parallel Distributed Processing. connectionism gained some kudos from being the revival of an approach that had been (allegedly wrongly) rejected in the 1960s, particularly following minsky and Papert’s analysis of the failings of perceptrons. It formed the basis of some interesting psychological models, particularly in the field of word processing, and related methods of computation found widespread application in the real world. connectionism is often seen as rejecting some of the fundamental tenets of cognitivism, such as discrete symbolic representations, and the need for structured relations (other than associ- ation) between them. Pure (functionalist) cognitivism is also sometimes seen as being undermined by the results of imaging techniques, which, it might be thought, make it foolish to propose theories of cognitive functioning that are completely abstracted from brain mechanisms.
another line of work that is sometimes seems as antipathetic to cognitivism in its original form is that based on the related ideas of embodied cognition, situated
cognition, and distributed cognition (including the use of external representations). traditional aI and cognitive psychology focus on processes taking place in individual minds, which they characterise as the manipulation of symbols that are usually, at least implicitly, conceived of as abstract. embodied cognition suggests that our concept of a chair, for example, is defined in part by the ways in which beings of a certain size and configuration (humans) interact with other objects (chairs) to perform such activities as working, eating, and resting. brain imaging techniques confirm that some of the same brain areas are active when we talk or read about chairs as when we use them. once stated, the basic facts of embodiment seem uncontroversial, and embodied cognition is best viewed as a development of cognitivism that calls on us to think carefully about the nature of our (internal) symbol systems and the mechanisms by which our symbols are grounded. likewise with situated and distributed cognition, it is hard to deny that some of our cognitive activities involve other agents and external representations. We do not usually play chess on mental chessboards, for example, but on real ones. and those boards store some of the information that we use in playing games of chess.
there are hard problems in correctly explaining both the embodiment and the distribution of cognition. However, neither set of issues poses a general challenge to the basis tenets of cognitivism. more radical challenges to cognitivism come from anti-representationalists, including some dynamicists and connectionists, and anti-empiricists, such as phenomenologists. In addition, John searle and some other philosophers argue that cognition is an essentially biological phenomenon, and that an abstract computational analysis is inevitably inadequate. Whether these challenges to cognitivism are decisive remains to be seen. Intellectual fashion aside, most, if not all, of the arguments for cognitivism remain valid.
Further reading
a detailed report of bruner’s work on concept formation is J. s. bruner, J. J. goodnow, and g. a. austin, A Study of Thinking (new york: Wiley & sons, 1956). n. chomsky, Syntactic Structures (the Hague: mouton, 1957) gives detailed exposition of an early version of chomsky’s syntactic theory. chomsky’s classic attack on the application of behaviourist ideas to language learning and language use his review of Verbal Behavior, by b. f. skinner, Language 35 (1959): 26–58. chomsky’s Aspects of
the Theory of Syntax (cambridge, ma: mIt Press, 1965) describes what came to be referred to as the
“standard theory” of syntactic structure within the chomskyan framework. a. W. ellis and a. W. young, Human Cognitive Neuropsychology (Hove, uk: Psychology Press, 1988) is a textbook providing an overview of the cognitive neuropsychological approach. H. kamp and u. reyle, From Discourse
to Logic: Introduction to Model-Theoretic Semantics of Natural Language, Formal Logic and Discourse Representation Theory (dordrecht: kluwer academic, 1993) is the standard text on the discourse repre-
sentation theory; and J. kimball, “seven Principles of surface structure Parsing in natural language,”
Cognition 2 (1973): 15–47, constitutes an early, highly influential paper on the issue of how surface
structure is computed in language comprehension. g. a. miller, “the magical number seven, Plus or minus two: some limits on our capacity for Processing Information,” Psychological Review, 63 (1956): 81–97, constitutes the classic paper applying information theoretic notions to analysing the capacity of short-term memory, and g. a. miller, e. galanter, and k. H. Pribram, Plans and the Structure of Behavior (new york: Holt, rinehart & Winston, 1960) sketches a generalisation of the information-processing approach to all aspects of psychology. m. minsky, ed., Semantic Information Processing (cambridge, ma:
mIt Press, 1968) is a classic collection of papers on semantic information processing from the mIt aI lab; a. newell and H. simon, Human Problem Solving (englewood cliffs, nJ: Prentice-Hall, 1972), an in-depth presentation of newell and simon’s information-processing approach to human problem-solving; u. neisser, Cognitive Psychology (new york: appleton-century-crofts, 1967), the first important cognitive psychology textbook. c. e. shannon, “a mathematical theory of communication,” Bell System Technical
Journal 27 (1948): 379–423, 623–56, would have to be considered the classic presentation of information
theory; and b. f. skinner, Verbal Behavior (new york: appleton-century-crofts, 1957), skinner’s attempt to apply behaviourist ideas to language learning and language use (as criticised by chomsky). the original presentation of cybernetics is n. Wiener, Cybernetics: On Control and Communication in the Animal and the