Stretching a point

This passage from Moore's notes on Wittgenstein's lectures bothers me:

If I want to find how elastic a rod is, I can imagine two ways: –

(1) With a microscope I can see the structure, & can say it is elastic.

But do I mean this structure by “elastic”? I might.

But (2) I might investigate by pulling the rod, & seeing what happens.

This might be what I mean, & the structure only a symptom.

So with “good”.

We might mean by “good” simply “action of this sort”…(Wittgenstein, forthcoming: May 9, 1933).

Specifically the line, "So with 'good'." What could this mean?

The interpretation I tried out before looked for the relative/absolute distinction from the Lecture on Ethics in this passage. But I'd like to try just to make sense of the passage on its own. The stuff about the rod seems straightforward enough: I might mean 'elastic' as a kind of technical term, saying something specific about the physical structure of an object. Or I might mean it in a more ordinary sense, in which it refers to the behavior of an object in certain circumstances. And we are talking about how elastic the rod is, so it's a matter of degree that concerns us.

How could assessing an action's goodness be like this? Nothing much suggests itself about what would correspond with the technical, structural sense in this case. My first thought is of ethical theories: an action might be called good if it promotes happiness, or is performed on the basis of a maxim that could be willed as a universal law, or accords with the Ten Commandments, etc. But it's not clear that examining the structure of the action (whatever exactly that might mean) would help here. Do the consequences, does the maxim, belong to the action's structure? Well, they might. Depends what you mean by 'structure' when the word is applied to actions. And that, I think, could be said about any suggestion regarding the correct interpretation of this part of the analogy.

So what about the pulling part? We can't really do anything to the action and see what happens. We could see what results the action produces, but it's unlikely that Wittgenstein believed that this is the way to judge goodness. A consequentialist reading of the analogy would also leave unexplained the pulling part and the reference to good's meaning 'action of this sort'. The word 'simply' just before 'action of this sort' is interesting too. Could good actions simply be those belonging to a certain type, not definable by their structure? That is, I take it, not definable by anything other than their belonging to that type? And where the type in question has a non-technical, ordinary name?

I think he must mean something like this. In which case I think the Lecture on Ethics reading is at least roughly right. Wittgenstein probably had Moore's open-question argument in mind here too. He was lecturing more or less to Moore, after all, and the definition/symptom distinction sounds relevant: whatever feature you might identify as the good-making property, it still makes sense to ask, "But is it good?" just as whatever feature is elastic-making one can still ask, "But does it stretch when pulled?" Wittgenstein is a non-naturalist, like Moore, although I think he would agree more with the autonomy-of-ethics thesis than with Moore's realism about normative judgments. Not because such judgments aren't objectively true or false but because the meaning of "objectively true or false" is far from clear in this case.

More to think about:

• the question of symptoms suggests a relation to questions about criteria, 
• Wittgenstein's early interest in Moore (along with Moore's view that our knowledge of moral truths comes from their being self-evident) suggests that On Certainty might have an ethical aspect/significance/reading, 
• Moore gives what might be called a metaphysical account of goodness whereas Wittgenstein seems to prefer something else (something more linguistic and more ordinary), 
• there are suggestions in all this of a connection between the notions of family resemblance and of secondary sense. 

This connection comes in two ways. First, the Lecture on Ethics both defines ethics in a family resemblance way and calls it nonsense in a way that Wittgenstein later associated with secondary sense. Secondly, the use of a word in a secondary sense depends on some felt but unidentifiable (is this right?) similarity with its use in its primary sense, and family resemblance is, or at least can be, like that. One kind of family resemblance says A is like B and B is like C, therefore C belongs to the same family as A despite their having nothing much in common. Another kind says A is like B despite their having nothing helpfully specifiable in common. Two shades of red, say. Sure they're both red, but in virtue of what? And if they count as the same color, why don't various shades of orange, pink, and purple? (In case this is all wrong let me distract you by pointing to more on family resemblances here.)