Searle claims that "language is essentially constitutive of institutional reality," (CSR, page 59) meaning that the institutional facts depend on language, in some way, for their existence. However, this seems to lead to a paradox, for Searle also wants to claim that language itself is an institutional fact, and it seems at first that if this is so, then we must also say that language itself depends on language for its existence. In this paper, I will explain first what it is about language that makes it an institutional fact, then explain the role it has in other institutional facts, and finally try to say what is special about the institution of language that makes it not dependent on itself.

There is one essential feature of language that ties it inextricably to institutional reality: that is, that it contains elements that represent, symbolize, or point to things other than themselves. These elements, in language, are words, and words have to have certain essential features. First, there is nothing in the physical nature of the words themselves which cause them to represent what they do; second, they instead do this representing by convention; and third, as a result they to this representing necessarily in a way that is understandable by a group of more than one person.

All these features are shared by what Searle terms "institutional facts". Analogous to words in language, there is nothing in physical reality that makes an institutional fact true or false; again, it is instead a matter of convention and public acceptance that allows these types of facts to exist at all. Most importantly though, institutional facts share with language this symbolic feature that gives something the ability to count as something else. It is this feature, as we shall see, that makes language constitutive of institutional facts.

Searle makes two important distinctions between "language dependent" and "language independent" facts, and then between "language dependent" and "language independent" thoughts. The former can be seen if we take a statement like "Mt. Everest has snow and ice at the summit." Although stating this fact may require language, Searle claims that fact itself has nothing to do with language, since it is true regardless of linguisitic conventions and regardless of whether a language exists to state it. On the other hand, a statement like "‘Mt. Everest has snow and ice at the summit’ is a sentence of English" is, according to Searle, language dependent, insofar as it is a fact that requires the existence of language to be true. The category of language independent thoughts in the second distinction covers considerably less ground, for only very simple thoughts like the desire for food, or other beliefs and desires that can be had by beings without language fall into this category. Many thoughts are instead language dependent, insofar as they require languistic elements just to be thought.

Let us examine in more depth what it is that makes a fact or a thought language dependent, for this should help us on the way to solving our paradox. Searle places two conditions on a fact being classified as language dependent: "First, mental representations, such as thoughts, must be partly constitutive of the fact; and second, the representations in question must be language dependent" (CSR, page 62) In other words, the fact has to have its truth value based at least in part on mental representations themselves, not solely on the things being represented. These mental representations, or thoughts, must not be of the type that are language independent, like hunger. Instead, the thoughts themselves must be language dependent.

Institutional facts, because they are based on constitutive rules, and rules are mental representations that have to be accepted and recognized, automatically meet the first of these requirements. Searle says that institutional facts are "ontologically subjective" but "epistemically objective", meaning that the fact that they exist at all depends on human mental processes, but then their truth or falsity can be determined with objective criteria regardless of these subjective processes.

But the question still remains, does this type of subjectivity in institutional facts make them language dependent? We cannot go the easy way and say that they are language dependent because we have words for these facts. In other words, it is simply not true that we need the word "money" to have money, since we can have practices involving money regardless of being able to talk about them. Instead, the language dependence stems from the structure itself of institutional facts, where, as Searle says, we have some thing X and we have it "count as" some other thing Y, in a context C. It is this "counts as" which should remind us of language. For just like in language, an institutional fact has something counting as something else not based solely on its physical features, but based on an assignment of function which gives the function because people agree to give it this function. As Searle says, "Its only money because we think its money!"

But to say that thinking that money is money makes the existence of money a language dependent fact, one has to make the huge assumption that thinking about money is a language dependent thought. After all, it seems we can use money without thinking about it, and without words. But the real question is, can we think that money is money, can we "count" certain bits of paper as money, can we assign a status function, without thought? Without language dependent thought?

Searle argues that we cannot, for two reasons, one practical and one logical. He asks first why any thoughts that are not explicitly about language are language dependent. Practically, we cannot have thoughts about large numbers and mathematical operations without symbols counting as those numbers. For example, it seems very difficult to think about 150 + 200 being equal to 350 without these symbols counting as these numbers. Logically, we cannot have a thought like "Today is Monday, March 30th" without words symbolizing dates. However, neither of these is an institutional fact, because there is no status function added by either these numbers or dates.

However, we can now begin to see that in cases where a status function is added, these types of symbols will always be necessary. When we have some X counting as Y, since there is nothing in the X term that prelinguistically allows us to see it as a Y. In other words, since the Y term only exists because of collective agreement, there must be some symbolic way we collectively understand the X as a Y. It is this symbolization that makes language constitutive of institutional facts. For example, if we take a game like football, which is an institution with constitutive rules, and we then examine one of these rules, like the scoring of points, we will see that the acts we count as "scoring" are essemtially linguistic, insofar as they are symbolic. In other words, there is nothing in the physical act of running accross a line with the ball which makes it count as a touchdown. Instead, just as we collectively assign meaning to certain words. we analagously assign meaning, or status, to the act of scoring a touchdown. The important thing to emphasize is that physically, the act of crossing the line with the ball is exactly the same as scoring a touchdown, the only difference being the symbolic, collective attachment of status to the physical facts.

In conclusion, we should now be able to see why language itself does not require language. In a word, the answer, as Searle puts it, is that "language does not need language because it already is language." (CSR, page 72) This sounds like a cop out, but it isn’t, since we have defined both language and institutional facts based on this ability we have to assign symbolic function to things. Language is just the most basic way in which we do this, with words counting as the things they mean. There is nothing in the acoustic sounds themselves, the sounds that we call words, which makes them mean what they do. Instead, we collectively agree to have them mean what they do. We then use this same technique with things like bits of paper and call them money. So in a way, it is not right to say institutional facts "require" language, they are merely a manifestation of the linguistic ability we have, which is most perfectly manifested in the fact that we have language, which is itself an institution.

April 1, 1997