Open Mike: Cornell OBI Conference—Chapter Four, Pragmatic information: Conclusion
|July 16, 2013||Posted by News under Cornell Conference, Informatics, Intelligent Design|
To facilitate discussion, we are publishing the abstracts of the 24 papers from the Cornell Conference on the Origin of Biological Information (here) at Uncommon Descent, with cumulative links to previous papers at the bottom of each page.
Here is the conclusion to John Oller’s chapter, “Pragmatic Information.” Oller is a professor of communicative disorders at the University of Louisiana.
Because of the series of combinatorial explosions that occur in progressing up the ranks in any layered hierarchy of representational systems, to find or generate any string that will qualify as a valid representation of any actual sequence of events in ordinary experience, or as a viable representation of any organism or any actual part of one, diminishes rapidly toward a vanishing point. Meanwhile, as the number of strings that are possible are exploding, the ratio of meaningful to meaningless strings at every level diminishes with each increase in the rank of signs and/or the length of allowable strings. As a consequence, the problem of finding (or generating) any valid (viable) biological strings by random processes is like the needle in a haystack problem magnified many times over. As Dembski, Ewert, andMarks  showed (this volume), the search for a needle presupposes a searcher. But the problem of randomly generating the searcher is vastly more difficult than any of the seemingly impossible searches we might hope for that person? robot? algorithm? to conduct.
But the difficulty does not end there. Linguistic analysis of natural language systems shows another profound problem, as was illustrated by Montañez, Marks, Fernandez and Sanford (this volume ). As valid (meaningful and viable) strings increase, the difficulty of generating them by stochastic processes rapidly increases. Also, as I have argued here, with each combinatorial explosion as we progress upward through sign ranks to their highest level, the ratio of valid strings to all that are possible diminishes toward a vanishing point with a numerator of unity and a divisor representing an uncountable multitude of multitudes.
In 1948, Claude Shannon proposed to measure information as the improbability of any particular message “selected from a set of possible [equally likely] messages” . He noted that “the messages” frequently “refer to or are correlated according to some system with certain physical or conceptual entities” which he referred to as “semantic [sic] aspects of communication” (p. 379) and which he setaside. In doing so, he conflated the abstract and general sort of meaning properly termed “semantic” (associated with generalized conventional Ss) and also the particular and concrete “pragmatic” content (associated with particular concrete Os — the actual persons, places, events, and the “syntactic” relations between them in space and time (the p-mappings). I suppose that the crucial meaning that Shannon set aside is precisely the kind connecting intelligible signs to the facts of ordinary experience — pragmatic information. I agree with what I understood Baumgardner to say in one of the early discussion sessions at the Cornell symposium: When talking about information we need to work with the sort of meaning that is distinctly “linguistic in nature” (also see Baumgardner 2009 ). I believe that we need to consider the dynamic character of pragmatic information as I have described it here. It seems to be as essential in biology as it is in linguistics.
One of the reasons, I think, that we tend to over-estimate our understanding of “our glassy essence” — and to underestimate the richness of the simplest signs — is that we tend to look right through the p-mapping of any valid S to its O. As the sign systems of a child come to maturity, the generality of the S reaches out very easily to signify all possible instances of the O greatly exceeding the relatively few actual instances that have been or will ever be encountered in a life-time of experience. The agreement attained between the valid p-mapping of any S to its O in a TNR thus achieves what Peirce referred to as the “unity of coherence”  — like a glove perfectly fitting a hand, or the bite when the upper and lower teeth fit together. The completed, well-formed-system, is a unified triad of the SpO kind. It enables the closest we can reasonably get, I suppose, to anything like “complete certainty” in the material world. Thus every TNR, though triadic in its internal elements, as a signifying unity singles out a stream of particular facts that are both distinct from all the rest and yet, by virtue of being a part of the whole material world, are connected with the rest of it and with all the other TNRs. As a consequence, they enable, as Pragmatic Information shown in the earlier analysis of child language development, valid generalizations beyond what is experienced. I agree with Edward T. Weinberger’s comment  that “a theory [Shannon’s] that totally ignored semantics was, in some sense, incomplete” (p. 105).
Weinberger went on to urge a definition of “ pragmatic information” in terms of “usefulness in making an informed decision” (p. 106). I would only want to generalize his approach to account for all intelligent judgments of any kind about the facts of experience. To me Weinberger’s most intriguing claim is that “the maximal amount of pragmatic information accrues to messages that engender complete certainty” (p. 109). In my linguistic approach to pragmatic information, a maximally informative representation would be the sort found in a name mapped onto a particular identity appearing throughout a faithfully reported true narrative. With a view toward measuring pragmatic information, we can say that it varies from a limit of meaninglessness at one extreme, near 0, to a limit of what seems to be the gold standard where the unity of coherence, near 1, is commonly achieved. Simple SpO mappings, at the foundation of valid representations such as we find in ordinary TNRs and in viable biological codes, exemplify the sorts that can be used to calibrate the high end of a scale of pragmatic information, and as I suggested, we can step down from there toward the lower end by degrees. More.
Note: All conference papers here.
See also: Origin of Biological Information conference: Its goals
Open Mike: Origin of Biological Information conference: Origin of life studies flatlined
Open Mike: Cornell OBI Conference— Can you answer these conundrums about information?
Open Mike: Cornell OBI Conference—Is a new definition of information needed for biology? (Chapter 2)
Open Mike: Cornell OBI Conference—New definition of information proposed: Universal Information (Chapter 2)
Open Mike: Cornell OBI Conference—Chapter Three, Dembski, Ewert, and Marks on the true cost of a successful search
Open Mike: Cornell OBI Conference—Chapter Three on the true cost of a successful search—Conservation of information
Open Mike: Cornell OBI Conference—Chapter Four: Pragmatic Information