usenet@xyzzy.UUCP (Usenet Administration) (05/03/89)
> I disagree [.. that discrete reaction of neurons to fairly abstract > objects in the visual field is "symbolic" ..]. > The significant point is that there is a *direct relationship* between > specific retinal cells and specific cortical cells which we feel we can > explain by means of what we understand of physics. [...] From: throopw@bert.dg.com (Wayne A. Throop) Path: bert!throopw Very well, Ray and I were indeed using "symbol" and "represent" differently. I'll adopt Ray's meanings for this posting. (I leave to another day the puzzling characterization of discrete representation as "analog".) Ray adds a further illustration of what distinguishes symbols for representations: > [..A..] direct [..causal..] relationship is what I mean by "represent". > The level > of the alcohol in a thermometer "represents" temperature. The "40" printed > next to the alcohol column "symbolizes" a quantity. Given this, I contend that computers represent the world around them every bit as much as humans do. Take for example the bit pattern stored in the usenet facesaver archives which represents my face. The magnetic spots on disk (or whatever) have a causal relationship to shape and visual properties of my face. Those magnetis spots represent me every bit as much as the patterns of firing neurons in some human's visual cortex represent me when that human looks at me. There is a causal relationship from me, through the camera, through various processing, ending in magnetic spots every bit as valid as the causal relationship from me, through an eye, through various processing ending up in (likely to be discrete) patterns in the visual cortex. Turning to other computer senses, we have the ascii representation of keystrokes which transmit my keystrokes to readers of this newsgroup. These are representations as Ray explains the term. Even the pattern of dots on your screen is a ('nother) representation of my keystrokes, because there is still a causal relationship. Symbolism only comes in when the reader assigns arbitrary (to the "physics" or "causality" of the situation) meaning to these bits upon reading the screen. Note that none of this shows any reason why sufficently complicated computer systems cannot be intelligent entities, capable of assigning symbolic meaning to abstract or concrete objects just as humans do. -- If at first you doubt, doubt again. --- William Bennett -- Wayne Throop <the-known-world>!mcnc!rti!xyzzy!throopw
GA.CJJ@forsythe.stanford.edu (Clifford Johnson) (05/04/89)
In article <5620@xyzzy.UUCP>, usenet@xyzzy.UUCP (Usenet Administration) writes: >> [..A..] direct [..causal..] relationship is what I mean by "represent". >> The level >> of the alcohol in a thermometer "represents" temperature. The "40" printed >> next to the alcohol column "symbolizes" a quantity. > >Given this, I contend that computers represent the world around them >every bit as much as humans do. To step further into the infinite regress of mirrors that comprise "authentic" forms of (direct) representation, I would further contend that a person can represent a computer -- and indeed does so when his actions are driven by rules based on computer output.