Machines

We might distinguish between a 'mechanism' and a 'machine'. 

A mechanism, as I want to characterise it, may be a discovered item - a set of interacting parts whose behaviour may be predictable if the relationship between the parts, and between the mechanism and its environment, is understood.

A machine is a mechanism accompanied by a promise: that it will function according to a normative rule.

This promise has to be a human promise, of course. A human being may construct a mechanism, but will only have constructed a machine if the constructor gives it a purpose and warrants (to some extent) that this purpose will be achieved.

This might seem to render all 'natural' mechanisms purposeless, by ruling out 'discovered' purposes. This isn't quite the case, however.

Human beings have always wanted to find a proxy for purpose in the natural world - this was provided by hypothesising deities (especially in the case of the 'argument from design'). The promise of evolutionary theory is slightly different, but still has a normative core:

Evolutionary theory gives an account of how a natural mechanism, including a 'human being' (as biologically characterised), can appear to have a purpose. The promise, here, is from the theorists - that if evolutionary theory is correct, we may rely on some of its products.

The tricky thing, of course, is that it appears to be the human 'biological machine' that attributes purpose to things. Evolutionary theory itself is only valid because it is intelligible when expressed in human language. Can we say something about the 'purpose' of this language?

Probably not. Partly because the 'machinery' of language cannot be identified independently from its normative and intentional concomitants ...

More on this later, I think.