Views from inside and outside

We decide what to count as participation in our conversation.  We decide what to count as 'language'.  So much depends on this decision that it may not feel like a decision, so it's worth explaining what I mean here.

No finite set of specific behaviours, described only as behaviours - i.e. without intentional content - could count as participation in a conversation.  Equally, no set of similarly described behaviours could be ruled out as a way of participating in a conversation.

If I hear someone calling me from another room, but investigate to find a recording device, I do no think that it was the recording device that was calling me.  If I speculate that, nevertheless, I was being called, and that the recording device was being used do to this, then I must speculate that someone was using it in this way.  To imagine that the recording device was calling me, I would need to imagine that it was much more than a recording device.

On the other hand, any distinguishable states of affairs - the presence or absence of a light in a window, a sequence of clicks, whether a clock is set fast or slow, could be used to give a 'signal', and so could be interpreted as having some semantic content.  Many things which look 'natural' - the disposition of pebbles on a beach, the entrails of a slaughtered chicken - can, and have been, given semantic content.  Sadly, on the other hand, many serious attempts to participate in conversations have been treated as incredible.

We cannot write down a set of rules for what can and cannot be interpreted as a signal.  We have lots of rules about the complexity of the signal that can be sent once we describe the possible states of the transmitting and receiving machinery (from Shannon), and we can write down rules for the transmission and receipt of valid signals given these descriptions, but we can say almost nothing about their semantic content.

We could not send a signal to an alien species which explained to them how they should interpret the signal we are sending.

Sometimes it seems almost impossible not to interpret some state of affairs, some 'signal', as having semantic content.   A literate person finds it hard not to read the words on an advertising billboard.  Perhaps you are finding it equally hard to resist reading what I am writing now.   This compulsion can feel almost mechanical, or empirical.  If we allow it to guide our search for semantic roots, we will, however, only find more or less complex signalling machinery.  We will never be any better off than the aliens.

On the other hand, I could not be explaining all this to you, nor you be trying to understand my explanation, unless you could attribute semantic content to what I am typing here.  You must take a certain normative attitude to this string of characters in order to count it as a participation in a conversation, and not as random, or as purely mechanically produced.  But you cannot be compelled to do this, which is why I think it's appropriate to speak in terms of making a decision.

The view of the aliens is the outside view - if they can 'see' anything at all it is the complex signalling system.  As Davidson might argue, however, it's hard to give sense to 'see' here without making them seem very like us, and so like possible interlocutors.  (The 'aliens' metaphor breaks down here - 'alien' is too human a description.)

Our view is the inside view - the view of the participants in a conversation.  And this argument is part of that view.  It is an exploration of the possibilities it provides, to see which are intelligible.  A proposed mechanical/signalling semantic theory would also be one of these possibilities - and so would fail as a fundamental account because such an account would be circular.  The rules which described the signalling system would only be rules by virtue of their role in the conversation within which the signalling system theory was being articulated.

We should also feel comfortable about setting aside any other deterministic metaphors here - these, also, can only be constructed within a working conversation, and so cannot be produced as accounts of its semantic content without circularity.  It is possible that any 'scientific' account that we can give of the world will turn out to be deterministic, but we cannot project that onto the conditions of possibility of giving such account, since these cannot be fully articulated.

It's worth adding something about the character of a potential interlocutor's 'decision' to participate.  Clearly this can't be a 'rational' decision in terms that can be represented within our conversation.  'I have considered the options and decided that it is better for me not to converse with you' is a move in the conversation.  If it is regarded as an occasion of exit, then it cannot mean what it might appear to mean if it was a conversational move.  It is automatically unintelligible.

It is perhaps better to think in terms of conversations being maintained or breaking down rather than in terms of participants joining or leaving.  The edge of the describable world is not on any map.