2. Look now at the useful notion of `customary complexes'. There is a sense in which all of our thinking about the world depends upon our identification of `customary complexes' (of states of affairs or of experience - we shall see later that these are interconnected) and our conditioning, thereby, to expect (i.e. assume) all the elements of a particular `complex' when we experience some of them. If the barking of a dog and the blowing of a whistle are, in our experience, part of a complex which includes the delivery of mail, we find ourselves going to the letter box to see what we have received whenever we hear those noises together.

a. We could not be conditioned into expectancies unless our experience of life were in `repetitive sequences' (customary complexes). The notion of recognising is, quite simply, that of seeing a situation (a sequence of experience - or of things experienced) as `the same again'.

It follows, then, that what I see as evidence that X pertains, when I am not actually experiencing X, is that set of circumstances which, together with X, form for me a `customary complex'. Thus, a certain `heaviness' in the air and black looming clouds overhead are, for me, evidence that a thunder storm is imminent (as indeed they are for my dog who is terrified of thunder and creeps under the table long before it actually starts). The `heaviness', the clouds and the flashes and crashes are elements of the same `complex'; when we observe some of them we expect (infer) the rest.

b. But, of course, not all clouds are (to me or to most other people) a sign (we shall look at signs shortly) of thunder - only certain kinds of clouds in certain circumstances. Here we are looking at the precision/vagueness relationship. Comparatively vaguely, it is a dog; more precisely it is a spaniel; still more precisely it is a cocker spaniel.... etc. Conceptually, complexes are `made up' of sub-divisions of complexes (and, being subjective, complexes are `conceptual entities'). In our identifications of states of affairs we move always, as it were, from the vaguer to the more precise sub-division. We see it as clouds in the sky (with whatever implications that has for us - and then, within that, as heavy dark clouds in the sky, black clouds accompanied by a `heavy atmosphere' and, at this point, the implication (our inference) is that it is a `thunder-storm situation' - i.e. we expect the flashes and crashes.

(c) We might say that two inter-connected things are happening:

i) We are noticing more and more detail and thereby reducing to (or `homing in on') more and more precise sub-divisions of our complex (sub-complexes).

ii) We are, thereby, expecting in much more precise and specific terms.

e.g. To see it as an approaching Toyota is to expect to see, when it gets closer, somebody driving it; to see it as her Toyota is to expect, when it gets closer, to see my wife driving it.

But note that these are inter-connected `happenings'; it is not that one occurs and then (and therefore) the other occurs, they are, in effect, just two descriptions of the same process; the more precise expectations are the identification of the sub-complex.

It should also be noted here that there is a direct relationship between the precision of our identifications of complexes (and, thereby, of our expectations) and the likelihood of error. As Bertrand Russell once put it very nicely, it is easier to hit the bulls-eye with a lump of putty than with a dart. If I identify an object as a tree, I am less likely to be mistaken (i.e. to find myself expecting experiences which I do not get) than if I identify it as a red-flowering gum tree.

Here think again about mistake-making. Any mistake occurs at the level of precision of complex-identification which is ` one step further than we should have gone'. [But note that this is merely analytic; making the error just is identifying the complex too precisely].

d. We can speak, however, only of the likelihood of accuracy, never of its necessity. We reason thus: This is an X type situation (customary complex) and all X-type situations include Y; therefore I can expect (assume) Y. This is logically justified only if Y-ness is included in X-ness by definition (part of what counts as X-ness) but, of course, it is not.

Here think back to `Tom is a husband, so Tom has a wife': If he is a husband, then he must have a wife - but he may not be a husband; being a husband is not part of what counts as being Tom.

Since every situation must be different in some way from every other situation (or they could not be two different situations - we shall look at this further under `Identity') it is always possible on any given occasion that the difference will be crucial to the prediction.

For instance, I identify the weather outside (as I write this) as thundery, as the customary complex which for me includes a storm about to start. But the Weather Bureau (which has more `evidence' than I have) has not forecast thunder today. So it could be that this weather pattern (customary complex) is like those I am identifying it with in all respects except that there is no thunder. But, then again, I allow that the Weather Bureau has more evidence than I have, but they cannot have all the evidence - nothing could count as this.

e. The trouble is that we think in terms of definitions whereas we actually identify (recognise) customary complexes (kinds of things or situations) by observation. Definitions (the kinds of things we find in dictionaries) are invariably by `genus et differentia' : they point out the `broad class' to which this kind of thing belongs and the characteristics which distinguish it from all other sub-classes of that `broad' class. Thus my dictionary tells me that a car is a wheeled vehicle; things could be vehicles and not be cars but, if they are vehicles and are wheeled then (by definition) they are cars. The example shows clearly enough how inevitably this kind of stipulation detaches itself from normal usage. Aristotle stipulated in this way that men are rational animals. If we took this seriously we would have terrible problems in identifying men; if rationality were taken to be merely a capacity for reasoning then most animals would count as men; if it were taken to be a capacity for what we consider sound reasoning about abstract matters, then idiots would not be men at all.

In fact, of course, we accept as a car anything that, in our view, is as much like the other things we accept as cars as they are like each other; we accept as a man anything as much (in our view) like other things we accept as men as they are like each other. And, in doing so, we tacitly acknowledge that there are not, and cannot be, any essential characteristics - merely a great many characteristics which are nearly always there.

f. Many people find that a useful approach to this is to think in terms of `any M of N'. There is nothing technical or mysterious about this; `N' happens to have been used traditionally to stand for any (unspecified) number - `M' happens to come immediately before `N' in our alphabet - so `any M of N' is merely any set of characteristics which is nearly all of the `total set'. Thus, if we were to list all the characteristics we would expect a man to have (and regard that as the `N') we would accept as a man anything that had `any M of' that N. One of the characteristics would, for instance, be two-leggedness - but a one-legged man is still a man.

The M, however, is always less than the N - so that `somewhere' there is what we might call `the M/N Gap'. It should be clear, however, that the M/N gap can occur anywhere - which is why there can be no essential characteristic of any class (kind, customary complex).

The implications of this for identifying customary complexes should be obvious. We recognise a situation as `one of those' and, thereby, assume the `full N' of what counts for us as `one of those' - but we do so on an `any M of N' basis. So - if our specific expectation happens to be in the M/N gap, then we shall not get what we expect. If we recognise Jones as a man, we expect him to behave `rationally'. If Jones happens to be an idiot we are, therefore, surprised by his behaviour. `Rationalness' happens to be in the M/N gap in this case - but we do not therefore say that Jones is not a man at all. [And here it is important to remember that the recognition is the expectations].

(g) To predict (to expect, assume, take for granted) is always to `go beyond the given'. This is an expression you will become accustomed to; it means to believe more than is actually presented to you, or implied by what is actually presented to you. In the terms of the present discussion what this amounts to is that we always see customary complexes in more precise terms than our present actual experience really justifies - and form our expectations accordingly. In short, we take the things we actually observe to be signs of other things which we are not actually observing now and, thereby, take these other things for granted. Thus all of our thinking about the world, our believing that certain states of affairs pertain, is in fact our `reading of signs' - i.e. our reacting to one situation as a sign of some another, or a wider, situation.



3. We need now, therefore, to examine the concept of signification.

a. The first point to note is that it has what might be termed a three-termed logic. If we say `Jones hops', that is a complete, intelligible statement. So we could say that hopping is `one-termed'. If, however, we were to say `Jones kicks', we would promptly be asked `What does he kick?' We can't kick without kicking something. So we might say that kicking is `two-termed'. And if we say `Jones gives' there are two further questions to be answered: what he gives and to whom he gives it. Giving is intelligible only as somebody giving something to somebody else. This is what we mean by `three-termed'. Since, as we have seen, nothing can be evidence in its own right, it follows that nothing can be a sign `in its own right'; for there to be signification, something must signify something else to somebody.

Thus we see that a sign is not an entity, it is a function. For Jones the jakaranda blossom may or may not be a sign of Spring (for me it was always a sign of exam-time!). Whether it is or not, it is still jacaranda blossom but only if it is functioning as a sign of Spring (if Jones observes it and thereby expects the other things he associates with Spring) is it a sign. Put very simply: A is a sign of B to C if, and only if, on perceiving A, C expects (assumes) B.

But note that we are not claiming that C says `By Jove, that A over there is a sign for B - so I'd better expect B'. Rather, he simply finds himself taking B for granted when he observes A. His experience to date (of customary complexes) has conditioned him to do so.

(b) One thing that is very difficult to avoid is thinking of states of affairs as neat, discrete little bundles - like it is raining now. States of affairs can be of any complexity and any state of affairs could be signified to anybody (it depends on his conditioning) by any other state of affairs.

Thus a particular pattering noise might be a sign to Jones:

i) that it is raining

ii) that the test match will be stopped (and therefore drawn) because it is raining

iii) that either it is raining or somebody is watering the garden

iv that it is probably raining

What is signified is just whatever he finds himself believing.

(c) Thus the same observable situation may be, for different people, a sign of quite different things. What does signify what to each individual is dependent upon his conditioning. We all find ourselves assuming that things will be as things were. So when we don't get what we expect - i.e. have misread a sign - we are naturally surprised and investigate to see what it was that was different in this case - i.e. why and how we had taken one (sub. class of) customary complex to be another. In this way we are constantly, and quite involuntarily, extending our `batteries' of customary complexes - and also re-conditioning our expectations, our inferences. Suppose that a small child has received sweets from a particular visitor on the three occasions he has visited. On the fourth occasion the child finds itself expecting sweets - but none are forthcoming. And the same on the fifth and sixth occasion. By this time the child will find that he has distinguished two different (sub) customary complexes (though he won't think of it in these terms!) - that man's visit when he is in a good mood and that man's visit when he is in a bad mood - and will be re-conditioned into not expecting sweets unless the man is in a good mood.

It is worth noting here that this re-conditioning by experience of failure (of misreading signs - i.e. expecting things that don't happen) is why it is that, although any of our beliefs could be wrong, very few of them actually are wrong.

d. It should be useful here to look at the notion of `mere coincidence'. We have all said at times `Oh no; it's just a coincidence' when two things have occurred regularly in sequence but we are denying that `there is any connection between them' - that is, we are denying that the first should be taken as a sign for the second. Suppose, for instance, that Jones's trouser-zipper jams occasionally and he notices that, whenever this happens, there is a sudden rain squall within//// ten minutes. He mentions this to Brown who says ` How odd - but it must be mere coincidence'.

Now this raises two very important questions:

i) How long will it be before Jones finds himself expecting rain when his zipper jams?

ii) What is Brown distinguishing from what when he `writes it off' as mere coincidence? What is he denying?

Look at the first - we do not choose to be conditioned (believing is wholly involuntary); we just find ourselves expecting X when we perceive Y. The first time it happens, Jones is most unlikely to assume any connection between the jammed zipper and the downpour - after all, there's nothing about a zipper to suggest a downpour. But then, there's nothing about anything (in itself) to suggest anything else (in itself) - except experience. We are dealing with possibilities, not with necessities. And, as Jones's experience mounts up, he cannot but find himself expecting the downpour; the circumstances are exactly parallel with all others in which he sees things as signs of other things. No doubt he will be puzzled that this happens to him and not to other people, and he will look for an explanation of this. But he'll still find himself reaching for his brolley when his zipper jams.

Imagine that some chap assured you that, whenever he sneezes, the number three horse wins the next race at Flemington - and you find, to your astonishment, that he is right. Would you not very soon start watching that chap for a sneeze and open a telephone account with a good bookmaker?

Our beliefs are what we find ourselves feeling, not what we rationalise. We tend to think of ourselves as, through experience, discovering the `natural order' of things. - but what makes it the `natural order' is simply that it does happen that way, not that it must.

So when Brown says that the zipper/rain sequence is mere coincidence, what he is saying is that it is not part of the `natural order', that, however many times it has happened like that, there are no grounds for supposing that it will happen like that again - i.e. that there is no customary complex which includes zipper-j amming and sudden downpours - i.e. there is no connection between the events.

Now, plainly there is (or has been) a quite clear temporal connection between the events - so what kind of connection is it that (so Brown claims) is not there? Obviously, a causal connection.

Here it is of no consequence whether Brown is right or wrong; the point to note is that to see A as a sign for B is to assume a causal connection between A & B. This, of course, is not necessarily to assume that A caused B; something else altogether (C) may have caused them both.

Later, under Causality and Explanation, we shall see that it is a misunderstanding to think of one event causing another event. But this does not alter the fact that all of our thinking (our believing, our sign-reading) rests on the assumption of caus al connections between situations.

e. Let's now look at some of the kinds of things we tend to think of as signs - because they happen to be (to function as) signs of much the same thing to most people. Consider:

i) very busy ants - a sign of a change in the weather

ii) STOP - a sign that traffic is required to halt.

iii the factory hooter - change of shift at factory.

iv) daffodils blooming - Spring has come (better weather etc)

v) a sore throat, aching back, temperature - having the flu.

v) GENTS - there is a lavatory in that direction.

vii) flags, badges, `logos' - membership of groups or organisations.

vii) smoke from a chimney - that a fire is burning.

Because we tend to respond in the same way as each other to these things, we tend to think of them as `signs in their own right' - but, as we have seen, there can be no such thing. To avoid confusion, it is as well to think of (and refer to) the actual observable itself - the busy ants, the hooter, the flag or whatever - as the sign-token. We then see that the same sign-token can be a sign of different things to different people: for Jones the factory hooter might signify knocking off time; for Mrs. Jones it might signify that a soap-opera is just about to start on the telly; for Brown, who works at the university nearby, it might signify heavy traffic jamming the road.

Note also that the same sign-token can signify (simultaneously - or according to accompanying circumstances - consider the implications of this) different things to the same person. The hooter on a Sunday may be a different sign from the hooter on Monday for Jones (or whoever) - and, indeed, there is an important sense in which it is a different sign-token. And for Brown it may well be both a sign that the traffic will be jammed and that there is industry very close to the university (and lots of other, conceptually quite independent, things too).

Making lists of `signs' like the one above is necessary for exemplifying and making certain distinctions, but it can be dangerous because it suggests that signs are special kinds of observables and it is essential that we dispell any such notions, that we realise that any observable could be a sign of anything to anybody.

Furthermore, every actually observed situation is a sign of something to the observer, since all of our awareness of the world about us is the reading of signs. As we shall see when we examine Sensation and Perception, we cannot but `go beyond the given' - i.e. see the `given' as the customary complex of which (for us) it is a sign.



4. It is useful, however, to make some distinctions - first between what we might call `natural signs' and `deliberate signs' and then within the `natural signs', between what we might call simple (or external) signs and what we might call symptoms (or internal signs).

(a) If we look at the list, we see immediately that the busy ants, the daffodils, the sore throat and the smoke function as signs (to Jones) only because of the way that Jones has been conditioned by his experience; nobody is trying to tell him anything - he is just making this identification of the natural order of things. But the STOP and the GENTS -, the flags and badges and the factory hooter have been deliberately created and displayed by somebody else in order to signify certain specific things to Jones and other people. These `deliberate signs' are what we call symbols and we shall examine these in the next section. It is important, however, to note at this stage that any symbol-token can be, and almost certainly will be, also a natural sign to its observer - but of a different, situation from that of which it is a symbol. To give a simple example - the flag flying over the car yard is put there to signify to people (as a symbol) that the yard sells Fords (or whatever) but it also signifies to most people (as a natural sign) that the wind is in a particular direction.

(b) Our immediate concern, however, is with the differences in natural signs. Here compare the busy ants and the sore throat - and allow that, for Jones, the first is a sign that the weather is about to change, the second is a sign that he is getting flu. Whereas there is no obvious connection between busy ants and bad weather (they seem to be two totally different things that just happen to occur regularly in that sequence), the sore throat is an integral part of the flu which it signifies. It is, we might say, one element of that customary complex; so, when we experience it we expect all the other elements. For this we can use the term ` symptom'; the sore throat, the sniffling and sneezing, the high temperature, the aches and pains are all symptoms of the flu. In the same way, the splash is a symptom of the fish-jumping complex - so that, when we see and hear the splash, we don't say `That was a splash'; we say `That was a fish jumping'. Perhaps the best (because the simplest) example is the iceberg. As we know, only a small fraction of an iceberg shows above the water - but no lookout man would ever shout `Tip of iceberg ahead!' For him, and all his shipmates, seeing the tip is seeing the iceberg. So think of a symptom as the bit that shows and is taken for the whole complex it is (presumed to be) a bit of.

(c) Here it is most important to note that there is not some sort of `entity' (separate from the symptoms) which has the symptoms (the way that a roof might have a coat of paint). The only `entity' is the complex which just is the totality of the `symptoms of it'

You can strip the paint off a roof and the roof is still ther e. But if you were to
`strip away' all the symptoms of X there would not be some kind of `bare X' left;
X just is `its' symptoms. Thus, whatever Jones sees as a symptom of X is, for Jones, a sign of everything else he sees as a symptom of X. In this way, the reco gnition of customary complexes is the `reading of' (response to) signs.

(d) Now consider again what we have called `simple' or `external' signs - like the busy ants signifying a change in the weather. It should be seen that the one does function as a sign for the other (to Jones) only because Jones sees (recognises) the ants as a symptom of a customary complex which also includes (as another symptom) a change in the weather.

Caution here - we are not suggesting that the busy ants are a symptom of the weather change (that would be plainly silly); we are saying that (for Jones) the ants are a sign of the weather change because (for Jones) both busy ants and changed weather are symptoms of a particular customary complex.

Here, think back to `mere coincidence'; if Jones did not see it this way then, for him, the fact that a weather change followed the business of the ants would be ` mere coincidence' - i.e. the one would not be a sign for the other.

But one further note of caution - we create our own customary complexes, and we do so by the process of reacting to signs. It is not suggested that Jones sees lightning and says `By Jove, that's a symptom of a customary complex. Now what are the other symptoms? Thunder. Rain. I'd better expect those.' He is simply conditioned by his experience to expect thunder when he sees lightning - and that conditioning is his thinking (i.e. reacting) in `customary complex' terms.

(e) Finally - think again about the `M/N gap'. We could now put it that what is recognised by Jones as an X complex is any M of the N symptoms which, for him, conjointly count as an X complex. Thus, the observation of any of these (M) symptoms (of the `N') is, for him, a sign for all the others. If the one which interests him happens this time to be in the M/N gap - then he will expect something which he does not get.

`Reading Signs' - i.e. seeing things as symptoms of complexes- is believing. The complex cannot be (since every complex must be different from every other complex - we shall look at this under Identity) precisely the same as the complexes experienced in the past - and may be different from these (the M/N gap) in a way that is relevant to the prediction (sign-reading) in question. From which it follows that the belief may be wrong.

This is not a different set of reasons why any belief could be wrong; it is merely a different description of the same set of reasons. To understand that clearly is to grasp the arguments we have dealt with up to this point.