12,080
edits
Changes
MyWikiBiz, Author Your Legacy — Wednesday May 01, 2024
Jump to navigationJump to searchDirectory:Jon Awbrey/Papers/Inquiry Driven Systems : Part 7 (view source)
Revision as of 15:27, 26 June 2013
, 15:27, 26 June 2013no edit summary
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Part 6|Part 6]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Part 6|Part 6]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Part 7|Part 7]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Part 7|Part 7]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Part 8|Part 8]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Appendices|Appendices]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Appendices|Appendices]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : References|References]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : References|References]]
====7.2.5. Discussion of Examples====
====7.2.5. Discussion of Examples====
At this point, in order to make further use of the pragmatic theory of signs without distorting its character too severely, I am forced, even for the sake of this simple example, to broach a difficult and potentially controversial topic that, for good or ill, can no longer be avoided. For future reference, let me entitle this issue as a question about "the general versus the restricted theory of sign relations". The problem for pragmatic theory and practical application is how to steer a middle course between the riptide of vague generalities and the narrow idols of reductive procedure.
In general, questions about the relationship of signs and interpretants, both inside and outside a computational framework, form a major concern of the pragmatic theory of signs, and will continue to be revisited throughout this discussion. Time and time again, questions about the true character of the interpretant role are found to abide at the heart of the problem, forming the cardinal points of investigation on which everything else hinges. The reason for this appears to be yet another brand of recursion or self similarity affecting the domain of inquiry. Namely, interpretants fill a role within their individual sign relations that is analogous to the role that sign relations used as formal models fill within the larger inquiry, providing a controlled mediation between the wide open domains of phenomena and language.
In the general theory of sign relations, which concerns itself with the formal properties of thinking processes, the interpretant domain is meant to encompass the full variety of mental impressions and affections: ideas, concepts, intuitions, intentions, impulses and dispositions to act, and every sort of intellectual construct, both cognitive and affective. This heady brew is not yet on the table for current consumption, but all of it imbues the aspirations of artificial intelligence with a measure of essential enthusiasm, and a spirit to which even those whose scope is focused on the objective dynamics of intelligent systems cannot turn down a glass.
Working within the constraints of formally effective descriptions (FED's), this project can explore nothing more than pale scattered shadows of these connotative and ideational dimensions of meaning. In the pragmatic theory of signs, it is held that all thought takes place in signs. Ideas, concepts, and other mental constructs are regarded as signs in the mind, in other words, as modifications of their peculiar medium that affect the states of their conducting agents. But with regard to their pertinent formal structure, namely, the sign relations that shape their action as signs, mental signs are no different from the generic brand of signs. In sum, all signs are defined as signs precisely in terms of their relative associations and formal operations, and not according to the marks of any absolute material essence.
Still, there falls within the sweep of this enlarged scope an element of crucial importance to all forms of pragmatic thinking. This is the notion that a solid store of meanings for signs and ideas can be found, not quite in the individual actions of an agent that come and go with the moment, so transient and irrepeatable in themselves, but in the agent's conative character, informed conduct, and consummate disposition to act. This realm includes both inborn and borne in patterns of inclination, broadly conceived plans of action, and generalized contingent resolutions to act in definite ways under prescribed conditions.
If this description sounds familiar, it is with good reason. The formal realm that shares these features is precisely the domain of operational definitions and effective programs whose use in clarifying concepts is being explored in this project. The upshot is, that whatever vague signs are allowed into the sign domain simply on the chance that they might mean something to somebody sometime, there is a more critical property demanded of the interpretant domain that is intended to play a role in deliberate inquiry. For a sign relation to serve inquiry in a positive way its interpretant domain ought to make available to its agent explicit expressions of program like entities that can define with maximal clarity the imports of its signs and ideas.
====7.2.6. Information and Inquiry====
====7.2.6. Information and Inquiry====
Given my initial description of inquiry as a process that reduces the uncertainty of an inquirer (any observant agent or interpretive system) about the state of an object system, and combining this with the characterization of interpreters in concrete form as sign relations, there arises an obvious question that must be addressed by this project: How is the state of uncertainty of an agent about an object system to be defined from the data present in a sign relation?
In spite of the bare construction of the A and B dialogue it is possible to elaborate a few scenarios on its basis that illustrate the relevance of sign relations to inquiry situations. To devise motivating stories for these inquiries and still be able to obtain the needed variations from such sparce materials, I will be forced to re use many elements of the sign relations A and B in non standard ways. Because they lack most of the analytic refinements that will be needed for complete clarity, these inventions risk the introduction of a few confusions. However, the exercise of untangling potential confusions in a simple example can provide useful practice, highlighting problems before they grow too complex to tackle, and training the attention to detect what features really matter in defining a situation of inquiry.
One type of inquiry might begin with A having no idea what B will say next, except for the certain knowledge that it must be confined to the syntactic domain S = {"A", "B", "i", "u"}. I will refer to this as a "syntactic" type of inquiry, since the object system appropriate to the inquiry situation, as described, is identical to the syntactic domain S. As a rough approximation, this inquiry can be viewed as a degenerate spin off from the original dialogue, one in which the true object domain has been lost, and attention has devolved to mere banter over signs. More carefully regarded, the relationship of the syntactic inquiry to the original situation could be described as deriving new sign relations Syn (A) and Syn (B) from the old sign relations A and B, in each case replacing O with S and splitting the stock of ordered triples in a corresponding fashion, as shown in Tables ** and **.
In this sort of "syntactic inquiry", the state of uncertainty on the part of A about the state of the object system S is a condition of maximum entropy with respect to the outcomes in S and can be represented as a uniform distribution of probabilities over S. In this scenario, A has log2|S| = log2(4) = 2 bits of uncertainty about what B will say next. If A hears B say "A" next, say, then A has no remaining doubts about the issue. As a result of receiving this sign, A comes to reside in a state with 0 bits of uncertainty about the question now past. The same reasoning applies to each of the other signs in S. Altogether, each sign in S conveys 2 bits of information to the interpreter A with respect to the prior condition of maximum uncertainty about the state of the object system S. The "average uncertainty reduction per symbol", in this case 2 bits, is called the "capacity of the information channel", as this channel is defined by the entire set up of the inquiry situation.
<pre>
Table **. Sign Relation of Interpreter X
Object Sign Interpretant
"A"A A "A"A
"B"A A "B"A
"i"A A "i"A
"u"A A "u"A
"A"B B "A"B
"B"B B "B"B
"i"B B "i"B
"u"B B "u"B
Table **. Sign Relation of Interpreter Syn (B)
Object Sign Interpretant
A "A" "A"
A "A" "u"
A "u" "A"
A "u" "u"
B "B" "B"
B "B" "i"
B "i" "B"
B "i" "i"
</pre>
Another type of inquiry might begin with A wondering what object B will denote next. Here, the object system referred to as a part of the inquiry situation is identical with the object domain O of the sign relations A and B.
'''Fragments'''
To deliver the logical functionality that is required to support inquiry, a computational framework must incorporate the ability to work with both empirical and rational knowledge. To do this it needs to have signs that refer to particular experiences and symbols that represent types of experience, and it needs, not only the capacity to examine the bearings of each upon the other, but a means to express the gist of this result in an integral form.
If these references and representations are to avoid the various ways of violating the bounds of sense — something they can do either by failing to have sufficient denotation from the outset or by exceeding the bounds of consistency and tractability at any stage of attempting to process their indications — then ...
# Operating on corrupt arguments or initially senseless indications. Attempting to start out from a state of empty nonsense, from a logically pointless or impoverished point of view, and trying to pursue a moment of semantic irreference on the impulse of a direction with null import. Drawing on resources that are logically empty and following instructions that are semantically nil.
# Transgressing the bounds of consistency or tractability at any subsequent stage of computation and thereby becoming logically empty or effectively vacuous, conceptually inconsistent or computationally intractable at an intermediate stage of investigation.
Even though the present discussion is focussed on isolated cases of sign relations, it cannot illustrate the properties of these examples in an adequate way without considering extended multitudes of other relations, both those that share the same properties and those that do not. Thus, to get the comparative study of sign relations started on a casual basis, something that is helped in addition by placing sign relations within the larger field of n place relations, I will exploit a few devices of taxonomic nomenclature, intending them to be applied for the moment in a purely informal way.
dimensions of temporal evolution and deliberate evaluation
coordinating temporal evolution with directed evaluation
partial specification: approximate, deficient, imperfect, incorrect
partial satisfaction: approximate, deficient, imperfect, incorrect
----
----
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Part 6|Part 6]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Part 6|Part 6]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Part 7|Part 7]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Part 7|Part 7]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Part 8|Part 8]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Appendices|Appendices]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : Appendices|Appendices]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : References|References]]
• [[Directory:Jon Awbrey/Papers/Inquiry Driven Systems : References|References]]
</div>
</div>
----
----
[[Category:Artificial Intelligence]]
[[Category:Artificial Intelligence]]
