Skywriting week 1

What struck me most about the Turing reading was the unimportance of the machine's physical construction. And what I found most interesting about the second reading was what it might imply for language learning as computational. Most of my thoughts were centered around the last reading. This paper begins by defining “physical symbol systems” (pss) as having necessary and sufficient means for intelligent action. The word symbol is defined, so is the symbol system and finally what a physical symbol system is, however, this keyword “intelligent” is not defined aside from as something which can describe the action of a PSS. The word action isn’t defined either, I’ll let that slide, but not without commenting on the claim that symbols must affect action. Is it not enough for a symbol such as a thermometer reading to be simply enjoyed and understood? Or are we considering the action potential which underline understanding as “actions” in themselves. Along that line of thought: I do not understand what is meant in lines 7&8 when it is claimed that symbols in a PSS are physical things contained in brains, is this referring to the neural readout from symbol recognition? Or symbol visualization?

Majorly, this reading made me reflect on the types of symbols and models that we use in science and in everyday life and how these symbols whether they be analogue or digital map onto what they are trying to model. Specifically, when the text claims that models are not judged based on correctness but usefulness it made me think of the Bohr model of the atom (this idea of undependable yet useful models maps onto the popular argument for realism). Lastly, and what I consider most interestingly, I could not help but wonder: is language a symbol for meaning? Or speech? Does the level of abstraction serve a similar purpose in models as it does language learning?