Brentano and Frege

Sandra LaFave

Franz Brentano (1838-1917) said intentionality is the characteristic feature of the mental. "Since intentionality is, Brentano claimed, an irreducible feature of mental phenomena, and since no physical phenomena could exhibit it, mental phenomena could not be a species of physical phenomena. This claim, often called the Brentano Thesis, or Brentano's Irreducibility Thesis, has often been cited to support the view that the mind cannot be the brain, but this is by no means generally accepted today." (Dennett, Oxford Companion to the Mind, "Intentionality")

Intentionality, or aboutness, is a relation. In logic, a relation, or a relational predicate, is a predicate that states a connection between or among individual things. Examples of ordinary relations include:

... is taller than ...

... is a friend of ...

We can fill in the ellipses with names of individuals. In logic, we symbolize a relation with a capital letter, and use small letters to indicate the names of the individuals in the relation. For example, we write "Bill is taller than Al" as follows:


where "b" stands for Bill, "T" stands for the relation "is taller than", and "a" stands for Al.

Ordinary relations can exhibit logical properties such as symmetry and transitivity. If a relation is symmetrical, then yRx follows immediately from xRy. For example the relation "is a sibling of" is symmetrical, since for all individuals, it holds in both directions: if A is a sibling of B, then B is a sibling of A. But the relation "is a brother of" is not symmetrical, since if A is male and B is female, A might be the brother of B, and not vice versa.

The relation "is taller than" is transitive, which means that if x is taller than y, and y is taller than z, it follows necessarily that x is taller than z. The relation "loves" is neither symmetrical nor transitive.

Statements of ordinary relations maintain their truth values when expressions with the same denotations are substituted. For example, if "Bill" denotes the same individual as "the President" and "Al" denotes the same individual as "the Vice-President", then if it’s true that Bill is taller than Al, it’s also necessarily true that the President is taller than the Vice-President.

With this background, we can now talk about a special class of relations: those that exhibit intentionality, or aboutness. They include

___ knows ...

___ believes ...

___ desires ...

___ hopes ...

where we fill in the "___" part with the name of an individual and the "..." part with a proposition, typically beginning with the word "that". For example, "Jack knows that Olympia is the capital of Washington," "Jane believes that Europa is a moon of Saturn." We very commonly use statements of this type in folk psychology, and in the social sciences. However, these relations do not obey the same logical rules as the ordinary relations described above.

As Gottlob Frege (1848-1925) pointed out, truth values of statements containing intentional predicates (such as "believes" or "hopes") do not remain the same when identically-denoting terms are substituted. I.e., statements of intentional relations are not logically well-behaved!

Let’s look first at a relation that is well-behaved. "Is taller than" is well-behaved, in the sense that truth value does not change when we substitute expressions with the same denotation. "The President is taller than the Vice President" has the same truth value as "Bill is taller than Al" if we agree that "the President" is just another way to designate Bill and "the Vice-President" is just another way to designate Al. Quine says relational statements of this sort are "referentially transparent".

Statements containing intentional predicates are another story. If John believes he was born during the Carter Administration, and John also (wrongly) believes that Carter was the 40th President (Carter was the 39th), then John does NOT believe he was born during the Administration of the 39th President – even though "the Carter Administration" and "the Administration of the 39th President" denote exactly the same time period. If it’s true that "Fred hopes to be elected President" and it’s also true that the President is an assassin’s target, it does not follow that "Fred hopes to be an assassin’s target." Do you see the problem? Quine calls it "referential opacity in intentional contexts".

In other words, you can’t do ordinary logic in intentional contexts because the ordinary rules about substitutability don’t apply! How, then, can you have sciences that rely heavily on intentional statements (e.g., economics, with its models of rational desire)? Remember that intentionality is the characteristic feature of the mental. Yet you can’t do logical inferences with intentional statements: you can’t pin down what a belief or hope is about, since the "aboutness" depends on how the subject frames the belief, within a whole web of the subject’s other beliefs. Faced with this problem, many philosophers of the early twentieth century, especially the analytic philosophers of the Vienna Circle (the logical positivists) voted to oust both the Mental and the mental from truly scientific discourse.

In psychology, the behaviorists were already working to rid the science of psychology of Cartesian Mind-talk (since there was no way to experimentally verify or falsify statements about invisible, insensible "things"). So the positivist philosophers gave the behaviorists yet another philosophical justification for purging psychology (and the human sciences generally) of any words that seemed to refer to the Cartesian Mental – all the intentional words like believing, desiring, hoping, feeling, imagining, preferring, etc. The behaviorists even got rid of "sensation" words such as "pain".

For a time, until around the 1960’s, many intellectuals agreed with the basic Behaviorist approach. But there were persistent problems. The behaviorist approach snuck intentional language in by the back door. The behaviorists didn’t allow experimenters to say "The animal is hungry," since of course you can’t directly observe hunger (hunger exists metaphysically subjectively, only as felt by a consciousness). Instead, you were supposed to say "The animal exhibits food-seeking behavior." But that formulation is still intentional, because seeking implies a seeker or seeking subject. "Seeking" is something an agent does, and whether or not an agent is seeking (as opposed to just meandering about) is a function of what the agent intends.

Gilbert Ryle (1900-1976) agreed with the anti-Cartesian physicalism of the behaviorists, but also agreed with Brentano that you can’t adequately describe the behavior of sentient beings without intentional language. Ryle formulated a clever compromise: logical behaviorism. Logical behaviorism insists that you talk about intentional states like belief and desire to explain behavior; in other words, logical behaviorism lets you invoke folk-psychological explanations. But logical behaviorism systematically translates statements of folk psychology into conditional statements. A conditional statement has the form

IF (antecedent) THEN (consequent)

Ryle then requires that the antecedent and consequent be publicly observable events in the world of the senses. So you get to have intentional descriptions of behavior without any spook stuff. How does Ryle do this? See notes on Logical Behaviorism.

Sandy's X10 Host Home Page | Sandy's Google Sites Home Page
Questions or comments?