British Ethical Theorists 0003: Thick Evaluative Concepts and the Is-Ought Gap

The previous post brought up the fact that the British Ethical Theorists currently under discussion all seemed to think that "thick evaluative concepts" (understood as those that are both descriptive and evaluative (e.g., "liar", "bigot", "generous", "courageous") can be reduced to a small number of basic "thin" concepts such as "good" and "right".


Philippa Foot vs. R.M. Hare

The International Encyclopedia of Philosophy article on thick concepts discusses a dispute between R.M. Hare and Philippa Foot on the possibility of deriving "ought" from "is".

Foot argues that this is possible, using the concept of "rude". Basically, Foot argues, "X causes offense by indicating lack of respect" is a purely descriptive statement. This purely descriptive statement implies "X is rude". "X is rude" is an evaluative term that implies "X is bad". In fact, it implies "X ought not to be done" which, if valid, would result in an instance of deriving an "ought" from an "is". Even without this final step it derives a value from a fact.

R.M. Hare responded by means of a counter example (a standard method of showing logical invalidity). According to Hare's counterexample. "X is a German" implies "X is a kraut". "X is a kraut" is an evaluative term - a term of contempt. Clearly, being a German does not imply that one is deserving of contempt, so we have an example here of true premises and false conclusion. An inference form that generates a false conclusion from true premises is invalid. Thus, by Hare's argument, Foot's argument form is invalid.

From the point of view of desirism, one notices that Foot's example provides an instance of "thwarting the desires in question." Foot's "rude" action is "such as to thwart the desires in question" - that is, of causing offense by indicating a lack of respect. Hare's counter example uses a term, "German", that does not imply "is such as to thwart the desires in question". For an advocate of desirism, this explains why Foot's case derives an evaluation from a factual statement and Hare's does not. Foot's example is "such as to thwart the desires in question" and Hare's is not.


McDowell's Disentangling Argument

This section contains notes from the Internet Encyclopedia of Philosophy article on Thick Concepts. I went there because Thomas Hurka's book British Ethical Theorists discussed how the reductionism of these British Ethical Theorists answered McDowell's objections against the reducibility of thick concepts.

First, to understand McDowell's argument:

Recall that thick concepts are both descriptive and evaluative. They describe a type of object of evaluation, and they evaluate that which they describe.

This is called the "disentangling argument" because it works by disentangling or separating the descriptive component from the evaluative component. If a disinterested observer were to come across a community using one of these thick evaluative concepts, they will be able to understand the "descriptive" component and pick out its scope. However, the scope of the evaluative part of the concept is somewhat different. This is the goodness or the badness of the thing being evaluated. The disinterested observer would not be able to pick out the scope of the evaluative component. Consequently, he would not be able to pick out accurately the scope of the descriptive and evaluative component behind.

Let me try to explain it this way. Take the term, "Liar". This has a descriptive component - somebody who asserts as true what she knows to be false. The descriptive component includes novelists and under-cover police agents. The evaluative component excludes these. An outside observer can understand the descriptive component of a term like "liar" and, in a sense, know that novelists and under-cover police agents are liars. But the observer would not be able to accurately capture the scope of the term "liar" unless the observer also knows the evaluative component - which liars are bad. From this, McDowell draws the conclusions that thick evaluative concepts are irreducible - the evaluative component is only knowable to those who feel the evaluation.

According to Hurka, the British Ethical Theorists have a response to this Then, the response to McDowell's argument. They agree with McDowell that the descriptive scope is not entirely accurate. Instead, it describes the general range of the term (e.g., liar = people who assert things they know are not true). The evaluative component uses a thin evaluative term (e.g., "good") to then restrict or point out a subset of objects that the descriptive term identifies that are good or bad. Consequently, they argue, a thick term can be reduced - but the reduction is to a descriptive scope that has a particular thin-concept value. A liar is somebody who asserts things that they know are untrue where those assertions are, in some sense, bad.


Desirism and The Disentangling Argument

As I described earlier in, "The Meaning of Evaluative Terms", evaluative terms have four components: (1) the type of object evaluated using the term, (2) the desires in question, (3) whether the object of evaluation fulfills or thwarts those desires, and (4) whether it fulfills or thwarts those desires directly or indirectly.

McDowell's disentangling argument disentangles (1) from the other three components.

Desirism will agree that the scope of (1) is different from the scope of (1) through (4). There is a difference between a knife and a good knife, or between a song and a beautiful song. Let us assume that there is a culture where, in its language, instead of using the phrase "good knife" to refer to knives that have good-making qualities, it uses a single word. We may call this word "goodknife". This will not change the fact that a goodknife is just a knife that is good. It will only change the language that is used to talk about such things. Similarly, we may replace the phrase "teller of untruths for bad reasons" with the term "liar". When we disentangle "teller of untruths" from "for bad reasons", we get a scope that is more restricted then we get from "teller of untruths" alone.

Where desirism would differ from other theories is that it holds that (2) through (4) are also descriptive. It is fully reductive. This implies that McDowell is wrong about one of the premises in his argument - the premise that states that a disinterested observer cannot understand the scope of the thick term. This outside observer can know the desires in question. The outside observer can know whether (or which) of the objects of evaluation are such as to fulfill or thwart the desires in question. The outside observer can know whether (or which) of the objects of evaluation fulfill or thwart those desires directly or indirectly. Thus, the outside observer can know the full scope of the thick components. She can know the descriptive scope captured in (1) and the restrictions to that scope introduced in (2) through (4).

In fact, knowing the relationships of states of affairs and the desires of people other than ourselves is extremely important. This is how we explain and predict their behavior. This is how we predict how they will react to the things that we do - to incentives and deterrence, to promises and threats. McDowell's claim that the outside observer cannot know the scope of thick concepts implies that outside observers will not know how to please or displease, offend or honor, the members of that community.

I can know what another person likes and dislikes without liking it or disliking it myself. I may not be able to know what it is like to like or dislike something in a particular way. I do not share my wife's tastes in movies - but I can pick out a movie that she will like. Though I may be a mistake from time to time, this fact does not prove that, no matter how much I know about her, I cannot even theoretically be able to correctly identify the movies she would like.

McDowell's disentangling argument is supposed to rule out reductionist theories of thick evaluative terms. I do not see that it succeeds.