The instructional information processing account of digital computation

internal state” (Piccinini 2007a, 514, italics added). The mechanistic “rule” also highlights this distinction. “A rule [...] is simply a map from input strings of digits, plus possibly internal states, to output strings of digits” (Piccinini and Scarantino 2011, 8). This suggests that digital computation proceeds from input D and possibly the internal states of the computational system to output D. We conclude that mechanistic digits qualify as D. On the IIP account, internal states of computational systems constitute sd at the appropriate LoA. For example, the internal state of the TM’s controller and the input datum – that is, the scanned symbol – play an equal role in determining the next move of the TM and together they qualify as sd. To what type of information do mechanistic digits rise? They are not just ud as can be inferred from Piccinini’s emphasis on the ordering of digits (Piccinini 2007a, 515). Since the ordering of digits matters, they are sd. Piccinini argues elsewhere that digits “need not carry Shannon information at all [... and] strictly speaking, computing systems need not process Shannon information” (Piccinini and Scarantino 2011, 27). He also argues that digital computation need not be the processing of nonnatural (i.e., semantic) information, which is descriptive and representational. So, digits need not carry nonnatural information (or be representations) (ibid, 30). Do some digits qualify as instructional information? If digits are sd and the IIP account is correct, it seems that the answer must be affirmative. This is where the mechanistic “rule” comes into play. Since some digits act as input sd and others are output sd, it makes sense that some digits encode the rules for processing input sd. “Both data and instructions are physically instantiated as strings of digits, which computers can manipulate” (Scarantino and Piccinini 2010, 326). Furthermore,