###### Inescapable Consequences, Inescapable Contradictions Armed with an understanding of argument structure, we can now return to our first goal: showing that Moravec's argument is incorrect. Let us start by simply taking Moravec's argument seriously on it's own terms. This means that instead of critiquing it with respect to our own world view, we taken on his world view and see where that leads us—we are investigating the *consequences* of his explanation. We can start by taking aboard his claim that we should accept any mathematically possible decoding. This means that we are discarding notions such as isomorphism and structure preserving transformations, and thus decoding is just an arbitrary computation. %%TODO: Need to include these in prevailing%% So we are granting his definition that removes part of the explanation about what encoding and decoding are. Again, that is fine for we are addressing his argument on *it's own terms* and seeing where it leads. We move forward accepting any mathematically possible decoding, of which Moravec correctly identified the most general would be a lookup table. But we then must account for another question: how do we come to possess this lookup table? It must be physically instantiated in order for us to interact with it. We cannot simply posit some abstract object and assume we have access to it. We need to have a procedure for constructing it. And this is where Moravec's argument—*on it's own terms*—gets into hot water. In order to generate the lookup table that maps states of one system, say the rock, to states of the simulation, you must *run* the simulation. But that means that you are not *identifying* the simulation within the rock—you are *projecting* the results from running the simulation elsewhere *onto* the rock! You are not explaining anything about the rock itself. At this point we see that the rock plays no role in the explanation of the simulation. The simulation has been run elsewhere in some other physical system while the rock just sat there. Moravec then added an arbitrary mapping, claiming that this showed the rock had performed the simulation. This violates Occam's Razor for adding unnecessary complications to the explanation. But the most devastating consequence of this inspection is we see that *by Moravec's own definition of simulation*, this does not count as a simulation! Recall that Moravec clearly defined a simulation as being entirely defined by it's intrinsic rules and the physical substrate that they are instantiated in. We can see that in this case the simulation occurred elsewhere and produced a descriptive artifact, the lookup table. But Moravec then wants to claim the rock is simulating a conscious mind—even though the simulation was run elsewhere. We can clearly see that his argument—*on it's own terms*—contains a contradiction. By his very own definition the rock is not simulating anything! This makes his argument an exceedingly bad explanation. It has created many problems and solved none. %%TODO: Reference description vs explanation?%% %%TODO: This may need to be linked back to the structure of his argument—based on the consequences implied %%