Section 4
The Weak and the Strong Universalizability
Let me distinguish the weak universalizability from the strong universalizability (although the words sound very similar, my distinction is quite different from Gibbard's 1988, 59-60; as I see it, his distinction is rather concerned with weights of preferences, which will be discussed in the subsequent sections 5, 6, and 7). I will agree with Hare that a descriptive or evaluative word (having a descriptive meaning) is universalizable in the form of (2), and I will call this the weak universalizability. And if we can eliminate all words containing a reference to an individual ( e.g., the Paris standard), or if we can replace all such words with properly or semantically universal words (i.e., words with no reference to individuals), I will call this the strong universalizability. Then, my point can be expressed in a word; namely, the logic of a descriptive word does not necessarily demand the strong universalizability.
If we grasp this point, the rest of my argument is quite easy. The existence of a criterion for a descriptive or evaluative word does imply the weak universalizability, but not the strong universalizability, because the criterion may contain a reference to an individual. The existence of a reason for a value-judgment does imply the weak universalizability, but not the strong universalizability, because the reason may contain a reference to an individual. Thus, if we wish to assert the strong universalizability of a value-judgment, we need more than the logic of a descriptive meaning, more than the logic of a criterion, more than the logic of a reason (I pointed this out in Uchii 1974, but I did not know Sidgwick well then). Thus, although Sidwick may not have known the modern logic, his intuition was quite acute. When he asserted his Principle of Justice is not tautologous, he was basically right. The strong universalizability of 'ought' or 'right' has some substantive content, not provable by logic alone.
There remain the following questions: Then, (a) is 'ought' universalizable in the strong sense, (b) and if it is why? But I will not pursue these questions here. All I wish to point out is that if you want to give an affirmative answer to the first question (a), you have to defend your answer on a stronger assumption than Hare's; the fact that many people admit the strong universalizability of 'ought' does not establish that it is a logical thesis. Some may wish to appeal to the concept of morality (e.g., by asserting that at least 'moral ought' is universalizable), and others may admit that the strong universalizability (with respect to evaluative words) is itself a substantive ethical principle, despite its formal and abstract character. But in either case, its justification is needed. Notice that, even if we make the universalizability true by vitrue of the meaning of 'moral', we thereby import another substantive question, 'why should we be moral?' Thus, although many of us are, unlike Sidgwick, unhappy with an appeal to 'self-evidence', Sidgwick's claim that the (strong) universalizability of 'ought' is non-tautologous seems still correct.
A moral statement, Hare says, is an overriding prescription that is universalizable: the prescriber must stand ready to prescribe the same thing no matter what position he is to occupy. . . .
. . . Weak universality requires only that I prefer all told the same alternative for any position I might occupy. It does not require that my preferences all told be equally strong for each of those positions. I can care what position I occupy, so long as I do not care enough to reverse the direction of my preferences all told. If, on the other hand, a person's preferences all told are position-independent in strength as well as direction, then I shall call them strongly universal.
To 3. Is the Universalizability True on Logical Grounds?To 5. Universalizability and the Concept of Good
[back to Sidgwick&Hare index][back to Uchii index]
July 20, 1998. Last modified April 17, 2006. (c) Soshichi Uchii
webmaster
