Equivocation in Feinberg’s “Infinite Regress” Argument Against Psychological Egoism

What may not be evident from the title of this post is that I’m fairly sure that there is an equivocation in Feinberg’s “infinite regress” argument against psychological egoism. This appears in Section C of his article “Psychological Egoism” in the popular anthology Reason and Responsibility (12th ed., ed. Feinberg & Shafer-Landau).

Feinberg makes a point to distinguish the following:

(Mere) Desire Fulfillment (DF): DF is simply the “coming into existence of that which is desired” (479). For instance, a desire for the obtainment of some object x is fulfilled when and only when x is obtained. (With DF, there need not be the requirement that the bearer of a given desire experience its fulfillment. Death-bed wishes can be fulfilled.)

Pleasure2 (P2), or Satisfaction: This is the type of positive feeling/s that one tends to experience upon getting what s/he desired (and, of course, being aware that the desire was fulfilled). That is, P2 is the pleasure that often results from and in virtue of DF.

P2 is not, Feinberg repeatedly emphasizes, DF. And though the term “satisfaction” can denote either DF or P2, Feinberg makes it quite clear that he uses it with the latter sense in mind.

The Argument
The following argument is directed against the P2 egoist-hedonist [i.e. one who holds that P2 is “the sole ultimate objective of all voluntary behavior” (483)]:

[The “satisfaction thesis” of the P2 egoist-hedonist], however, is so confused that it cannot even be completely stated without paradox. It is, so to speak, defeated in its own formulation. Any attempted explication of the theory that all men at all times desire only their own satisfaction leads to an infinite regress in the following way:

[1] “All men desire only satisfaction.”
[2] “Satisfaction of what?”
[3] “Satisfaction of their desires.”
[4] “Their desires for what?”
[5] “Their desires for satisfaction.”
[6] “Satisfaction of what?”
[7] “Their desires.”
[8] “For what?”
[9] “For satisfaction”–etc., ad infinitum.

In short, psychological hedonism interpreted in this way attributes to all people as their sole motive a wholly vacuous and infinitely self-defeating desire. The source of this absurdity is in the notion that satisfaction can, so to speak, feed on itself, and perform the miracle of perpetual self-regeneration in the absence of desires for anything other than itself. (483)

And later on…

On the other hand, if [the egoist-hedonist] means pleasure2, then his theory cannot even be clearly formulated, since it leads to the following infinite regress: “I desire only satisfaction of my desire for satisfaction of my desire for satisfaction . . . etc., ad infinitum.” (483)

The Problem
As this argument is directed specifically at the P2 egoist-hedonist, one would hope that Feinberg invokes the concept of P2 somewhere in the argument. Presumably he does at lines 1, 5, 9, and so on, and takes himself to do so throughout the whole argument. The problem is that the language of lines 2, 3, 6, and so on (specifically, the “satisfaction of” locution) only makes sense on the DF reading of “satisfaction.” If we replace “satisfaction” with a definition of P2, I think this becomes more apparent:

[1′] “All men desire only pleasant feelings resulting from and in virtue of DF.”
[2′] “Pleasant feelings resulting from and in virtue of DF of what?”
[3′] “Pleasant feelings resulting from and in virtue of DF of their desires.”
[4′] “Their desires for what?”…

Pleasant feelings are not of desires. Nor can we charitably replace “of” with “from,” for even that is not right. The pleasant feelings do not simply come from the desires but from their fulfillment.

I believe that Feinberg equivocates in this argument. Of course, if someone sees a charitable interpretation on which he is not, please share.

Further Thoughts
Based on some discussion with Steve Darwall and Peter Railton, I suspect that the spirit of Feinberg’s argument may still be preserved, though I have not worked out, to my own satisfaction (a third sense!), exactly how this will be done. It would involve a regress caused by the type of desire in which the P2 egoist-hedonist trades. What do you say, dear blog-post reader?

Also, were I a psychological egoist, P2 egoism-hedonism would not be my brand of choice. (Would or has anyone ever endorsed this brand of psych. egoism?? I have doubts about that.) Could it really be that our apparent desires for certain strong “physical” pleasure sensations (Feinberg’s P1) are only disguised desires for weaker P2 pleasures? Before rejecting psychological egoism on philosophical grounds, I would like to see a sophisticated version presented. Such a version does not make an appearance in Feinberg’s article, as far as I can see.


6 Responses to Equivocation in Feinberg’s “Infinite Regress” Argument Against Psychological Egoism

  1. dtlocke says:

    How about we drop the word ‘satisfaction’ and rewrite the regress as follows:

    [1“] All people desire only the feeling that they get when their desires are fulfilled.
    [2“] When their desires for WHAT are fulfilled?
    [3“] Their desires for the feeling that they get when their desires are fulfilled.
    [4“] When their desires for WHAT are fulfilled?
    [5“] Their desires for the feeling that they get when their desires are fulfilled.
    [6“] When their desires for WHAT are fulfilled?

  2. Steve says:

    Very nice rewrite. That’s roughly how someone else suggested to me that it be done and how I think it probably could be reworked.

    I do have some lingering questions that perhaps you or others have thoughts on:
    -Why exactly is the regress a harmful one?
    -Is any given desire going to be for its own DF-satisfaction or for some other desire’s DF-satisfaction?
    -How does this picture fit in with our normal conception of desiring some state of affairs–e.g. my (fervent) desire to hula-hoop in twenty minutes time? Can I simply not have a hula-hooping desire, given [1″]? Surely the P2 egoist-hedonist must have some story of this apparent desire (for we have tons of them), and I’m wondering what it would be.
    -Besides thinking that P2 egoism-hedonism is already a funky little version of egoism probably not worth defending, I worry that, even so, this funky view is not presented in its best light, if the regress argument succeeds. If it succeeds, doesn’t it do so because we literally have no other desires but those for DF-satisfactions? But where does hula-hooping fit in? Are there not instrumental desires? (Feinberg does sketch the position as holding that P2 is the sole ultimate objective of all voluntary behavior, so perhaps he thinks that there are.) Also, what individuates desires on this P2 egoism picture?…
    -A seed of a thought on what P2 egoism might involve: We want/desire certain states of affairs but, as it happens, only really care about one type of aspect–namely, the DF-satisfactions that can be reaped. Would the regress argument still go through in that case?

    Well, I’d be interested to hear what people think.

    [If no one wants to expend mental energy on an untenable position, I won’t be surprised or unsympathetic. If they do, perhaps we’ll learn some things about the nature of desire.]

  3. dtlocke says:

    “-Why exactly is the regress a harmful one?
    -Is any given desire going to be for its own DF-satisfaction or for some other desire’s DF-satisfaction?”

    Take some agent A and her desire D. Since the P2 egoist holds that all desires are desires for the feeling one gets when his desires are fulfilled, it follows that A’s desire D is a desire for the feeling A gets when her desires are fulfilled. But is D one of the latter desires? If so, we have a neverending circle. For now A’s desire D is a desire for the feeling A gets when A’s desire for the feeling A gets when A’s desire for the feeling A gets when A’s desire for the feeling A gets when… ad infinitum… is fulfilled. If not, then A has some desire D` among the latter. So now A’s desire D is a desire for the feeling A gets when A’s desire for D` is fulfilled. But, again, by P2, A’s desire D` is a desire for the feeling A gets when A’s desires are fulfilled. But is D` among the latter. And here we go again. At each step, we can either run around a circle or postulate a new desire. So we’re either going to get a circle or a regress.

    But is the circle/regress viscious? I think so, because it seems that our desires must be “grounded” in a way that they could not be if the circle/regress obtained. Maybe it has to do with our having finite minds, which simply couldn’t have such “infinite” desires. But although I think it’s clear that the circle/regress is viscious, I agree that is’t not EXACTLY clear WHY it’s viscious.

    As I was walking through Angell today I was actually toying around with a way to state psychological egoism that would avoid the cirlce/regress. Maybe we can do it as follows:

    P2`: For any agent A, there is a set of states of affairs S, such that for each member m of S, A desires only the feeling A would get if m were to obtain.

    Notice that the theory makes no mention of A DESIRING the members of S to obtain: rather, it simply says that A would get a feeling if a member of S were to obtain and A desires that feeling. The motivating idea behind moving from P2 to P2` is that what got P2` into trouble is that according to it A desires only the feeling A would get if A’s DESIRES were fulfilled. P2` avoids this (and thus the circle/regress) by just talking about the states of affairs directly, and making no claim to their being DESIRED.

    But, as you say, what of our usual talk of, e.g., desiring to hula-hoop? No doubt the psychological egoist will either say that we are systemicatically confused about what it is we desire and that what we really desire is the feeling we get when we hula-hoop or else paraphrase our talk of desiring to hula-hoop into talk of desiring the feeling we get when we hula-hoop and saying that the latter gives the real semantic content of the former. Of course, each of these strategies is implausible. Hence, psychological egoism: bad; hula-hoop: good!

  4. Steve says:

    I’m also sympathetic to trying to find a better formulation of the P2 thesis. I’m not sure the P2′ solution works though since P2 feelings must be distinguished from other types of feel-goods (e.g. P1 feelings). So, say S is the state of affairs of my getting to indulge in the consumption of some mint chocolate chip ice cream. Feinberg is going to want to say that the pleasure that comes from and in virtue of getting what I had wanted (P2) is one thing, the gustatory sensations that I get from that sweet mint-chocolatiness (P1) another thing.

    So one might ask of your P2′:
    -“What feelings are those?” or “What feelings do you mean?”
    -“The feelings that they get when their desires are fulfilled.”
    -“When their desires for WHAT are fulfilled?”

    Does that sound right? The idea is that you need to isolate the P2 feelings and may not be able to do so without referencing desires, which then leads to the circle/regress.

    It seems like we might get a solution if we allow for a host of instrumental desires, which are not “ultimate.” The ‘D’ in a given DF-satisfaction desire would then be some or other instrumental desire. …But, I’m not yet sure whether this makes good sense.

    Here’s an instance of a desire-pair:

    D1: The desire to eat ice cream
    D2: The desire for pleasant feelings that come from and in virtue of D1 being fulfilled

    (I should re/emphasize that these pleasant feelings will *not* include the gustatory pleasures that come from eating the ice cream. That is what Feinberg means by P1, and he doesn’t ever suggest that these classes overlap.)

    The formulation of D2 doesn’t lead to a circle/regress, right?

    Of course, how do we explain the origination of D1? Given its allegedly instrumental nature, it seems to natural to think it develops because its fulfillment is perceived to be a means to the fulfillment of D2. But D2 seems to require D1 for its own existence, and so couldn’t precede it. Can D1 arise first? It’s supposed to be a purely instrumental desire. So how could one of these desires come into existence before the other?

    Hmmm.. Maybe they just spring up together automatically given the appropriate perceptual cues. I see ice cream–boom, I’ve got D1 and D2. Interestingly, the dependency relation is what we might expect. If D2 disappears (e.g. because I learn, for whatever reason, that I will not experience DF-satisfaction), then we can expect D1 to disappear as well, ceteris paribus. And if D1 disappears, D2 might still actually survive. Say that I can get two DF-satisfactions if I don’t act on D1; D2 might still exist. Well, it might get altered. It might become the desire for the pleasure that would result from and in virtue of a desire that I did and might’ve still had but don’t now.

    The general idea is that desires tend to come in pairs, only one member of which is “ultimate.” And this might avoid the regressive problem. (This is not to say that P2 egoism isn’t blatantly false.)

    Well anyway, this is all pretty half-baked, so I’m sure it is rife with problems. Dustin, curious to hear what you and others think.

  5. Steve says:

    p.s. A conversation with Konek inspired the thoughts in the last comment. Well, only the good ones.

  6. Orcus says:

    In the limited extent that we choose to fulfil our desire this is a truism.

    Considering that P2 is effectively uncorrelated to DF and that on any event following P2 that future DF is shaped by out conclusion of P2.

    The problem shifts to one of pre-determinism of P2 from DF. Or paraphrased what are the external influences that will shape P2 in spite of or to spite DF.

    So you invent an other in a bite sized term, believe it is The external influence and seek to proscribe meaning to it. Having meaning we the shape our DF to reach P2.

    Retrospectively our environment shapes our expectation of P2 given our actions which in turn we may assign the The external influence as the creator of our condition. or historic P2.

    If I have reached any conclusion is that P2 cannot be predetermined or prepared or mitigated but it can be certainly be limited in its repetitions by change.

    Beyond a semantic game in a vacuously limited situation the debate seems circuitous.

%d bloggers like this: