What is it for someone to have unconscious opinions?

1. Let us start by looking at an example. John and George, who are not philosophers, are having the following discussion. John tells George: "So you believe that you should avoid making friends with male homosexuals." George replies: "I have never thought that." John continues along the same line, trying to convince George that, although he never had the thought that he should avoid making friends with male homosexuals, his behaviour exhibits this opinion. For this, he may point at several situations when, on meeting a male homosexual, George has avoided getting too personal with him. "This behaviour of yours", John says, "expresses the opinion that you should avoid making friends with male homosexuals." Eventually, George says: "Right, so I had the opinion that I should avoid making friends with male homosexuals." This example, I think, may be enough for some to start asking questions like: "What are unconscious opinions?" "How do they differ from conscious opinions?" and so on. For now, allow me to resize the problem to fit the time I have left for my talk. The question I will consider is: "What distinguishes George's unconscious opinion that he should avoid making friends with male homosexuals from his regular conscious opinions?"

2. Before taking a closer look at this problem, I want to note that there are people for which the above scenario is not a plausible one.

On one hand, the Cartesians would hold that something that is not accessible by introspection for a subject couldn't be attributed to that subject¹. If I cannot find out by introspection that I have a certain opinion, then it is not in my mind. Thus, for such a Cartesian², the above scenario is unacceptable. George should not have agreed that he held the respective opinion unconsciously. There cannot be unconscious opinions, or unconscious mental objects of any sort. I suppose that for a Cartesian rejecting this scenario the first-person belief attribution prevails over the third-person belief attribution. In addition, the first-person belief attribution is probably defined like this:

I have the opinion that p IFF I remember myself having thought that p in the past, toghether

with a feeling that I was accepting that p.

Now, I do not want to discuss about this position right now. I think that the above scenario is plausible. As such it is also functioning as a counterexample to the necessity of the Cartesian condition³.

On the other hand, we have the eliminativists about beliefs⁴. I suppose that they will say nothing against the example per se, but they will deny our talk about beliefs, either conscious, or unconscious. The reason for this, I think, is that they do not see cognitive processes as related to manipulation of sentences. A different reason could be this⁵. Belief attribution has not only the normative assumption that the subject is reasonable, but also requires evaluation of beliefs, which can be done only on normative grounds. Therefore, since the naturalist approach should be descriptive, if we are going to be naturalists about knowledge, we should give up speaking about beliefs or opinions. Again, I do not want to start a controversy about eliminativism regarding beliefs now. I accept the example and I think that one could speak of unconscious opinions starting from it. For now, I just wanted to point out that there are people for which my talk would be meaningless from now on. I only hope that those less inclined to use such philosophical intuitions at this point will follow me well.

3. Now I want to consider briefly a few attempts to give an answer to our problem. First I will consider some ontological approaches of beliefs, then some epistemological approaches. In the end I will provide my own answer.

From an ontological point of view, the answer should probably consist either in saying that George's opinion is a different kind of object from his regular opinions, or in saying that, while being the same sort of objects, the different opinions have different properties, or at least different relations to George.

Thus, I do not see how Frege could answer our question, since he says that all the beliefs are propositions, i.e. abstract objects, and he is not very clear about the relation between the subject having a belief and the respective belief, seen as an abstract object. In addition, it seems evident that neither the syntactic, nor the semantic properties of propositions are enough to distinguish between conscious and unconscious propositions.

Fodor's view that beliefs are sentences in the Language of Thought physically instantiated as neural processes in our brains does not seem very helpful either. I never observe myself or someone else speaking in the Language of Thought. From this point of view, all my opinions are unconscious. Again, the relation between a subject and these private sentences is not at all clear and the semantic or syntactic properties of such sentences are of no use for our distinction.

Generally speaking, it is not at all clear how an object could be called conscious only because of its sort, some of its properties or some particular relation with other objects. The ontological approach, it seems, either makes all the opinion conscious (as it was the case with Descartes), or makes them all unconscious. In fact, I do not see how the sort of an object, any of its properties or its relations with other objects could render it conscious. This goes not only for propositions (abstract objects), regular sentences, sentences in Mentalese, neural processes, but for Cartesian mental objects as well. For this reason, I will leave aside other views according to which beliefs or opinions are some sort of objects, for now.

The epistemological approach seems to be rather centered on belief attribution. Let us see what could provide us with an answer to our question here.

The folk psychology view holds that we attribute beliefs to someone by interpreting that person's behaviour in view of a common sense theory of psychology which contains generalizations about relations between psychological states (beliefs, desires, fears) and relations between these states and inputs from the environment or between these states and the subject's behaviour. On this account, there is no difference between the way we attribute opinions to others and the way we attribute opinions to ourselves. So, in a sense, there are still no conscious opinions. It is only an accident that it was John the one who rightly attributed the opinion about homosexuals to George, and not George himself. We are inclined to say that George's opinion was unconscious because he did not attribute it to himself. But why should it matter who makes the attribution, since there is no difference between first person attributions and third person attributions? Shouldn't we say, if we are to agree with this view, that the opinion does not become conscious after being attributed to the subject, no matter by whom?

The 'Simulation theory'⁶ rises other problems. It is not clear how undergoing a so-called 'belief-forming' process or simulating one would lead someone to belief attribution. It may lead to asserting a sentence, but not necessarily with the intention to express an opinion. And if we want to say that the verbal behaviour of uttering p expresses the opinion that p, and thus to attribute the opinion that p to the subject we are considering, we must take a further step, beyond the simple simulation.

To conclude, the only way in which a theory about the belief attributions could be able to make the distinction between conscious and unconscious beliefs seems to be by distinguishing between first person and third person attributions. For now I cannot see how this could be done without bringing introspection back into the game. But then the first-person belief attributions and the third-person belief attribution might be two completely different processes, with no relation to each other. So one might ask: "Why are both called 'belief attribution'?"

...............................................................................................................................................................

4. Let me look now at some useful distinctions. Uttering a sentence is not always expressing an opinion, for the simple reason that not all our sentences are assertions. More than this, we could assert something without believing it. So asserting a sentence does not always amount to expressing an opinion. We could say that p with the intention to suppose that p and see what does follow from it, to pretend for a moment that p, to ask whether or not p and so on. Is it possible to say something that we believe is true without expressing the respective opinion? George might say 'The door is open', believing that the door is open, with the intention to make his guest leave. I do not think that in this case he is expressing the opinion that the door is open. His opinion has nothing to do with his intentions while saying 'The door is open'. We might continue along this line for a bit. A judge says 'George is guilty' and he also believes that George is guilty. Still, the point is not what the judge believes at present. The judge pronounces a sentence. We would not usually say that he did express his opinion at that moment. It is clear by now, I hope, that I take expressing an opinion to be a sort of a speech act, different from inviting, pledging, informing and so on. The explicit act of expressing an opinion would most probably look like this:

"I believe that p"

"I have the opinion that p"

"I think that p"⁷

On the other hand, it is obvious that when I say of someone else that she believes that p, I do not express her opinion, but I attribute an opinion to her. I could attribute an opinion to myself, of course, which is different from expressing an opinion of mine. Expressing opinions and attributing opinions could be regarded as two completely different speech acts. In addition, I could attribute an opinion to someone else considering that something the other person says or does expresses that particular opinion. This is a different use of the phrase 'to express an opinion', since the intentions of the other person do not matter in this case. And of course, since asserting a sentence is not necessarily expressing an opinion, I could say "I believe that p" without expressing the opinion that p. To sum up, I think it could be useful to distinguish between:

'S expresses the opinion that p' / 'S says that p' / 'S says "I believe that p"

'S expresses the opinion that p' / 'S pretends that p' / 'S predicts that p'/ 'S informs that p' a.s.o.

'S expresses the opinion that p' / 'S's (verbal) behaviour expresses the opinion that p'

'S expresses the opinion that p' / 'S attributes the opinion that p to himself (herself)

...............................................................................................................................................................

5. Now, let us look at a very simple example related to our initial scenario. George avoids Smith. On seeing this, John concludes that George's behaviour expresses the opinion "I should avoid Smith" and attributes to George the opinion "I should avoid Smith." At the same time, George does not avoid Smith with the intention to express the opinion that he should avoid Smith. Neither does he attribute to himself the opinion that he should avoid Smith. After having a discussion with John, George agrees that he had the opinion that he should avoid Smith. How is this possible? George is also willing to attribute opinions (and perhaps even intentions) to himself by interpreting his behaviour. He does this all the time. So what is the difference between George's unconscious opinion that he should avoid making friends with male homosexuals and his conscious opinion that, let's say, his name is George? George has perhaps expressed the opinion 'My name is George' in the past, either to someone else, or to himself. In any case, he has attributed this opinion to himself at a certain point in the past. No one has pointed out to him that the opinion that his name is George could have been attributed to him even before his own attribution (and perhaps that he could have agreed with this attribution even then). Therefore, having this in view, he could say that he has the conscious opinion, or, more simply, the opinion that his name is George. In the case of 'I should avoid making friends with male homosexuals', George has accepted that he could have attributed this opinion to himself in the past even without having expressed it, based on his behaviour, and decided to attribute this opinion to himself for that time (even before expressing it at present). Perhaps the source of our puzzle was the difficulty to understand what George meant when he said, after his discussion with John, something like: "Right, so I had the opinion that I should avoid making friends with male homosexuals in the past." We are used with "I had the opinion that my name is George in the past", which is used both to express an opinion for oneself for the past and to attribute an opinion to oneself for the past. But we were not so familiar with an utterance used to attribute an opinion to oneself for the past without expressing it for oneself for the past. I believe that this was our problem and that the only possible solution can come from a better understanding of our language, with no need for an ontology of beliefs or of a theory about belief attribution.

6. In the end I want to remind you that I have tried to focus on one case only. There are other cases that could perhaps rise a problem that could be as well called "the problem of unconscious opinions". To name only a few:

S knows the right answer to a question without accepting it.

S could compute a very large sum, although he never made that computation before.

S acts as if he believes that he is the most intelligent person in the world, although he would never agree to that.

S acts as if he believes that "All objects fall down", but he cannot use generalizations.

S usually believes that all emeralds are green, without thinking it right now. (He is asleep right now.)

Someone attributes to S the assumptions of his view.

We could attribute to S the sentences entailed, conventionally implied, conversationally implied or presupposed by what S is saying, although S has never said those sentences, not even to himself.

German speakers attribute to S, who is a monolingual English speaker, the opinion that "Die Schnee ist weiβ."

I want to suggest that there is nothing wrong with speaking about unconscious opinions with respect to at least some of the cases enumerated above. However, if we are going to find ourselves at a loss, the solution should be provided by a better understanding of our language. This is an opinion for which I hope I have offered a small support in this presentation.