American Philosophical Quarterly, Volume 17, Number 2, April 1980: pp. 81-101.

ON REASONING ABOUT VALUES

WILFRID SELLARS

I

1. In spite of strong currents to the contrary, perceptual metaphors -- intuition, contemplation, insight -- continue to dominate our conception of logical and mathematical knowledge. Knowing is knowing that, and to know that is to see. Even when metaphors from the language of practice are allowed a place in the sun, as when we speak of constructing numbers, propositions and proofs, the elements are seen to fit and the outcome is a structure to be, if not seen, at least grasped.

2. Obviously there is good sense in these metaphors. Logical thinking not only inquires but, on occasion, achieves. There are, as James put it, perchings as well as flights. Contemplation in the ordinary sense may not be the pure, unchanging act of the perennial philosophy, but it is at least a perceptual consummation and is fitted for its metaphorical role.

3. One sees that contradictory propositions cannot both be true. And, in particular cases, this and other insights guide our thinking. Is this because we have a, perhaps innate, perhaps instilled, dislike for the illogical? Is it a contingent matter of fact that we prefer "p or not-p" to "p and not-p"? Or is it, perhaps, a synthetic necessary truth that the awareness of inconsistency arouses disbelief and stills desires.

4. However the fact is to be understood, it seems clear that there is a real connection between knowing logical truths and being disposed positively to believe the implications of our beliefs and, negatively not to think logically incoherent thoughts.

5. Of course, any such connection must be a matter of "other things being equal" propensities, for many different factors enter into the explanation of a course of thought. But if the awareness of logical truth did not, as such, exert pressure on the mind and contribute to shaping thought, it would he only per accidens that we were rational.

6. These considerations can be expanded to include another dimension -- one which, I believe, provides a vantage point from which to get an over-view of our problem. For what is involved is not only the awareness of logical truths, but the having of logical concepts and the understanding of logical words. These, too, imply the above propensities. Bluntly put, to understand such words as "and," "or," "not," "if," "then," "all," and "some" is to be disposed to conform (ceteris paribus, of course) to the simpler principles of elementary logic.

7. It is obviously true that the meaning of a logical word is a function of its contribution to the implications of sentences in which it occurs. That this is true, at least in part, of the meaning of every word may be less obvious, but is certainly of equal importance.

8. Most words contribute specific implications pertaining to specific subject matters. Thus the word "bachelor" in

Jones is a bachelor
brings it about that this sentence implies
Jones is unmarried.
And the attributive "large" in
Mickey is a large mouse
brings it about that this sentence implies
Mickey is large for a mouse.

9. There are however certain words which, like logical words, contribute general kinds of implication to the sentences in which they occur. Thus, consider the present tense copula in

There is thunder.
It brings it about that the sentence implies
There will have been thunder.
To "understand the meaning" of these words is to "see" these implications, and to have the corresponding propensities pertaining to inference and belief.

10. To "see" that

'P' implies 'Q'
is, ceteris paribus, to have the propensity to believe 'Q' if one believes 'P.' But since, if 'P' implies 'Q,' "not-Q" implies "not-P", it is equally to have the tendency to believe "not-P" if one believes "not-Q,."

11. Thus to "see" that 'P' implies 'Q' is, in part, to have the propensity not to believe both 'P' and "not-Q" -- which leaves open the following combinations of belief in 'P' or belief in "not-P," with belief in 'Q' or belief in "not-Q"

believing 'P' and believing 'Q'
believing "not-P" and believing 'Q'
believing "not-P" and believing "not-Q,."
Which of these combinations is realized is not, of course, determined by "seeing" the implication.1
1 If we also take into account the fact that to "see" that 'P' implies 'Q' is also, in part, to have the propensity to believe 'Q' if one believes 'P' and to believe "not-P" if one believes "not-Q" then we must also take into account the combinations in which there is an absence of belief with respect to either 'P' or "not-P" or an absence of belief with respect to either 'Q.' or "not-Q." Of these 12 possible combinations only two conflict with the propensities in question, namely, (a), "believing 'P' and not-believing 'Q,'" and (b), "not-believing 'not-P' and believing 'not-Q.'" Which of the remaining combinations is realized is, again, not determined by "seeing" the implication.

12. For example

"X is made of lead" implies "X is heavy for its size,"
so that "seeing" this implication, I will have a propensity not to believe with respect to a certain object, O, that
O is made of lead and O is not heavy for its size.
which leaves open for belief,
1. O is made of lead and O is heavy for its size
2. O is not made of lead and O is heavy for its size
3. O is not made of lead and O is not heavy for its size.
Which of these beliefs I arrive at will be a function of other factors, e.g.
  1. I scratch O and it looks like lead. I will tend to believe in the first alternative.
  2. I heft O and it doesn't feel heavy for its size. I will tend to believe in the third alternative.
  3. I scratch O and it looks like lead; I heft O and it doesn't feel heavy for its size. I will tend to believe in the second alternative.2

2 Of course, if I scratch O and it looks like lead and heft O but it doesn't feel heavy for its size, I just might be on the road to a revolutionary disconfirmation of the implication itself.

13. On the other hand, I may have no encounter with O, and no evidence pertaining to its composition or specific gravity. In this case if I have any belief about O connected with "seeing" the above implication, it will be that

O is not both made of lead and not heavy for its size.

II

14. The examples I have chosen so far have come from discourse directed toward what is the case, as contrasted with what ought to be the case -- discourse about facts as contrasted with values. In this connection we have considered

  1. subject matter independent implications which accrue to sentences from logical words.
  2. subject-matter pervasive implications which accrue to sentences from words pertaining to spatial and temporal location.
  3. implications which are tied to specific subject matters, as in the example just discussed.

15. This classification, highly informal, makes no claim to be either tidy or complete. Indeed the second category was introduced to prepare the way for a further category of subject-matter pervasive implications -- implications which accrue to sentences by virtue of the fact that they contain expressions of a kind to which I have not yet directed attention, but which will be at the center of the stage for the remainder of this essay.

16. Consider the word "would" as it occurs in

Would that Jones were here!
It is a word which has a very general function. Sentences involving it, as we say, express a wish. In using it, one isn't describing an object. Rather, one is, in a suitably broad sense, expressing an attitude towards some conceivable object or state of affairs.

17. What are the logical or, should I say, conceptual implications of this word? What implications does it contribute to sentences in which it occurs?

18. In the first place, a person who candidly or sincerely says

Would that Jones were here!
can be expected, other things being equal, to try to bring it about that Jones is here. Thus, if a person comes to think it possible to bring it about that Jones is here they3 will -- other things being equal -- attempt to do so.

3 In this essay, to avoid the awkward 'he/she' I shall use the somewhat idiosyncratic but by no means non-existent construction 'a person ... they' and 'one ... they.'

19. Thus one who sincerely says

Would that Jones were here!
and believes that by doing A, Jones's presence can be secured, will, other things being equal, do (or try to do) A.

20. This, of course, raises the question, What are the 'other things' which must be 'equal'? In a sense this question is the topic of the argument to follow.

21. Let us try the following ideas on for size:

  1. The person must believe that there is a real possibility that Jones will be present.
  2. The person must believe that by doing A they will be doing something which contributes to Jones's presence.
  3. The person must have no desire which overrides their wish that Jones were here.

22. The first is connected with the difference between a wish and a desire.

The second is connected with the difference between a desire for a state of affairs and the intention to do an action which might bring it about.

The third is connected with the difference between what one wishes, some things considered, and what one wishes more things considered.

Let us take up the latter two in order.

23. We began our discussion of practical implications with wishes, which, as such, are conceptually remote from action -- even though they have a conceptual tie with action. As we saw, other things being equal, "x wishes that-p" implies "x tries to bring it about that-p." Let us now turn our attention to purposes which are very close to action, thus

I will take the 4 PM flight to Houston today,
where this expresses an intention, and is not a mere prediction. Compare
Jones will take the 4 PM flight to Houston today.

24. In the former context "will" is a cousin of "would that." Its implicative tie with action is to be found in the fact that, ceteris paribus,

Jones candidly says "I will do A in 10 minutes"
implies
Jones does A in 10 minutes
which involves the idea that, ceteris paribus,
Jones candidly says "I will do A right now"
implies
Jones does A right now.
Here "ceteris paribus" guards against the possibility that, for example, Jones is paralyzed or about to be struck dead, and presumes that A is an action which Jones is in a position to do.

25. Thus we can lay down the principle

Ceteris paribus, people carry out their here-now intentions
which is a special case of
Ceteris paribus, people do what they intend to do
and this, in turn, of
Ceteris paribus, people realize their purposes.

III

26. A bit of linguistic regimentation is in order. The above use of "will" -- the use in which it expresses a purpose or intention -- is limited in Standard English to first person action contexts, thus

I will do A.
But we can, of course, have intentions which pertain to the obtaining of states of affairs other than our own actions.

27. Thus it can be my purpose that my children have a good education. Obviously I can wish it:

Would that my children have a good education.
But if I think it possible that this state of affairs might obtain, I can desire it. Interestingly enough the Standard English way of expressing this desire is by the use of the auxiliary "shall," thus
My children shall have a good education
where this, as before, is no mere prediction, but expresses the intention or purpose that this be the case. To achieve a uniformity of representation, i.e. to avoid a constant shuffle between "will" and "shall," I shall use the auxiliary "shall" in all my examples to represent the expression of an intention. After all, it is so used in connection with most of the kinds of intention in which I shall be interested; and, furthermore, "will" in Standard English serves for the most part to form the future indicative.

28. Thus, instead of

I will do A
I shall use
I shall do A
to bring it in line with
It shall be the case that my children have a good education.

29. Furthermore, since I am not particularly concerned with mere wishes (i.e. with cases in which one is "favorably disposed toward a state of affairs" but doesn't think that it can be realized, I shall concern myself with the logic of intentions, and hence with the logic of my regimented "shall."

30. The most fundamental principle of this logic is a reformulation of an earlier point about "will." It is to the effect that

Jones candidly says "I shall do A here and now"
implies, ceteris paribus,
Jones does A here and now
Put more abstractly, the principle affirms a connection between
having intentions to do A here and now
and
doings of A.

31. The next logical feature to be noted concerns the connection between the statements

My children shall have a good education
My children will have a good education only if I put aside money
and
I shall put money aside.

32. Before examining the logic of this sequence in depth, let me make an obvious, but nevertheless crucially important, point about intentions. It is essential to distinguish between the autobiographical factual statement

I intend to do A
which ascribes an intention to me, as
Jones intends to do A
ascribes an intention to Jones, and, on the other hand,
I shall do A
which (in our regimented idiolect) expresses that intention.

33. Notice, however, that in the first person case the functions of the two sentences overlap. Thus

I intend to do A
not merely ascribes an intention to me, but also, for reasons familiar to students of Moore's Paradox4 serves to express it. In the third person case, thus
He, Jones, intends to do A
the statement obviously does not express Jones's intention to do A. I can express an intention about Jones doing A by saying
Jones shall do A
which would express my intention that Jones does A" and is equivalent to
It shall be the case that Jones does A.

4 I.e. The logical oddity of "It is raining, but I don't believe it."

34. It is therefore essential to distinguish between what is implied by an intention and what is implied by the fact that a person has the intention -- just as it essential to distinguish between what is implied by a belief (i.e. the proposition believed) and what is implied by the fact that a person has the belief (believes the proposition).

35. Let us equate for present purposes

Jones believes that-p, e.g. that it is raining
with
Jones, if in a candid frame of mind, would assent to "It is raining" (or its counterpart in Jones's idiolect).
and correspondingly
Jones intends that-p be the case, e.g. that his children get a good education
with
Jones, if in a candid frame of mind, would assent to "My children shall get a good education."

36. Thus we can distinguish between the implications of the sentence

It is raining
e.g.
The streets are getting wet
and the implications of the sentence
Jones believes that it is raining
e.g.
Jones expects to get wet, if he ventures outside
and, correspondingly, between the implications of the sentence
My children shall get a good education
e.g.
My children shall go to college
and the implications of the sentence
Jones intends that his children get a good education
e.g.
Jones believes that his children are docile.

37. Notice that if one proposition 'P' implies another proposition, 'Q,' it doesn't follow that a person who believes 'P' also believes 'Q,' but only that they are subject to criticism on logical grounds if they believe 'P' and either don't believe 'Q,' or believe "not-Q,."

38. Similarly if one intention implies another intention it doesn't follow that a person who has the former also has the latter, but only that they are subject to criticism on logical grounds if they have the one and don't have the other or have the contradictory intention.5

5 Notice that in most contexts in which I refer to intentions, the latter are to be construed as intention candidates or intendibles. The context should make it clear whether I am discussing a had intention or a mere intendible. Roughly, intentions stand to intendings as propositions to believings.

39. Thus, the intention expressed by

"I shall eat a healthy diet"
implies (by virtue of the principles of nutrition)
"I shall eat complete proteins"
but a person may very well have the former intention without having the latter.

40. This example calls attention to the fact that just as one proposition may imply another without a person realizing that it does, so one intention may imply another without a person realizing that it does.

41. Of course, when it comes to actual cases of reasoning, the only implications that enter the psychological picture are the implications which the reasoner takes into account. But from the standpoint of logical criticism and evaluation, the implications are what they are whether or not the reasoner takes them into account.

IV

42. With the above distinctions in mind, let us return to the main line of the argument.

43. We were interested in the special implications which attach to the word "shall." Of these the most basic was that which relates the intention expressed by "I shall do A here and now" to action. We can now see that this implication relates the having of such an intention, to action. Thus

Jones has the intention to do A here and now
implies, ceteris paribus,6
Jones does A here and now.
Also
Jones intends to do A in 10 minutes
implies, ceteris paribus,
Jones will do A in 10 minutes
(if, e.g. he doesn't change his mind).

6 The qualification is obviously necessary to take into account, for example, the possibility that the actual circumstances prevent Jones from doing A, or that he unwittingly becomes paralyzed.

44. From these implications we must distinguish those other implications relating to "shall" which concern the intentions themselves and not the fact that someone has them.

45. Thus, consider

My children shall get a good education
or, in general,
It shall be the case that-p.
Intuitively, this implies.
I shall (ceteris paribus) do that which contributes to bringing it about that my children get a good education.
The "ceteris paribus" is clearly necessary, although its exact role remains to be discussed. The crucial point, however, which it helps to highlight, is that there is a conceptual tie between
It shall be the case that...
and
I shall do ...
Abstractly put, intentions that something be the case imply intentions to do.7

7 It will shortly become clear that the distinctive feature of such implication lies in the connection between it shall be the case that I do A' and 'I shall do A.' The remaining features of the above example, including the ceteris paribus clause, reflect the implicative pressures which relate one 'shall be the case' to another, i.e. which transcend the distinction between 'shall be' and 'shall do.'

V

46 With one exception to be noted in a moment, and one possible exception to be considered near the end of the argument, these are the implications which pertain directly to "shall" and specify what might be called its conceptual grammar. Expressions of intention have other kinds of implications, but, as we shall see, they are grounded in implications between indicatives.

47. The fundamental principle connecting indicative implications with the logic of "shall" is

S-Imp.: if 'P' implies 'Q,' then "It shall be the case that-P" implies "It shall be the case that-Q".8
It is because of this dimension that the logic of purposes (and values) is largely derivative from the logic of facts.

8 This principle can be formulated as a biconditional, but I forbear to do so lest I open the door to misunderstandings, based on other things I have written, which there would be no time to alleviate.

48. The fundamental simplicity of this encompassing principle enables me to cover a substantial amount of ground by means of a few regimented examples. But first a word or two about the nature and scope of the principle are in order.

49. In the first place, implication is to be understood as a relation which authorizes inference and, more generally, serves as a criterion for the compatibility of beliefs and, correspondingly, of intentions, along the lines discussed in the opening sections. We noted there, without attempting a complete classification, that implications may be logical in the narrower sense -- subject matter independent -- or they may arise from categorial features of the concepts involved, or, as in the case of causal connections, they may be tied to specific subject matters. The latter clearly play a central role in practical reasoning.

50. In the last case it is important to distinguish between causal implications which are explicitly general and can be formulated as lawlike statements, and causal implications, which, while presupposing the truth of some lawlike statement, do not specify these statements -- which may not even be at hand. Thus we may know that in circumstances like these, 'P' causally implies 'Q' without knowing exactly what it is about these circumstances by virtue of which the implication obtains.

51. The task of botanizing implications and the ways in which they depend on other items for their existence and their knowability is, as I have implied, a difficult one. Indeed, it is, to a surprising extent, unfinished business. Fortunately, it will be sufficient for my purposes to rely on our intuitive sense of which matter-of-factual propositions imply which other matter-of-factual propositions, and concentrate on the relation of such matter-of-factual implications to practical reasoning.

VI

52. Another framework point remains to be made which is fundamental to all that follows. It is a point which has seemed to many whose judgement I respect9 to fly in the face of reason. The implications between intentions which arise from implications pertaining to matters-of-fact concern only the content of intentions and not their status as intentions.

9 I have particularly in mind Bruce Aune and Hector Castaneda, who have repeatedly taken me to task on this matter.

53. This thesis is the upshot of the distinctions which were drawn in sections III and IV above. It is reflected in the fact that, as I shall argue, negation, for example, enters into the expression of an intention only as part of what is governed by the "shall" rubric.

54. Thus, in our idiolect, we have

I shall do A It shall be the case that-p
I shall not do A It shall not be the case that-p10
but not
I not-shall do A
It not-shall be the case that-p
e.g.
My children not-shall have a good education.

10 For our purposes, "It shall not be the case that-p" is equivalent to "It shall be the case that not-p." Problems which concern the relation between "It is true that not-p" and "It is not true that-p" which arise in connection with the distinction between the Principle of Bi-Valence and the Principle of Excluded Middle have no bearing on the problem to be discussed concerning a possible double occurrence of negation in expressions of intention.

55. If we regiment

I shall do A
as
Shall [I do A]
and
It shall be the case that-p11
as
Shall be [p]
then we would have negations
Shall [I not do A]
and
Shall be [not-p]
but not
Not-shall [ I do A]
nor
Not shall be [p].

11 Note that the construction "It shall be the case that-p" takes advantage of the logical equivalence of 'p' and "It is the case that-p" to permit "shall" to exhibit its role of auxiliary verb, which "Shall [p]" does not do.

56. To generalize this claim, before submitting it to closer scrutiny, logical operators do not occur outside expressions of intention, or, to put it in other words, expressions of intention do not occur within the scope of logical constants and quantifiers. Thus we do not have

Shall be [p] or shall be [q]
as contrasted with
Shall be [p or q]
nor
For all x, shall be [if x is a future danger, then I will escape x]
as contrasted with
Shall be [for all x, if x is a future danger, then I will escape x].

57. This conceptual point is not a consequence of the fact that expressions of intention are neither true nor false for, as I pointed out in "Thought and Action,"12 they do have the semantical values, realized and not-realized depending on whether the correlated expression of belief is true or false. The fact, however, that they have these semantical values does nothing to counter the idea that an expression of intention as an expression of intention can only be negative by virtue of having a negative content. "Not-shall [ I do A]" is false in the radical sense of conceptually incoherent. The fact, when it is a fact, that my intention to do A is not realized does not make the (nonsensical) 'not-shall [I do A]' true.13

12 In Keith Lehrer, ed., Freedom and Determinism, New York, 1966.

13. We shall see, towards the end of the argument, that there is a place for another pair of values for expressions of intention, namely reasonable and unreasonable, but this neither requires nor permits an external negation.

58. It is time, now, to take into account a fact which seems to make nonsense of the above claim. I have in mind the obvious fact that expressions of intention can occur in a text along with other expressions of intention and, indeed, with declarative sentences which are expressions of belief.

59. Again, is it not a fact that

I intend to do A and I intend to do B
makes perfectly good sense? If so, should not
Shall [I do A] and shall [I do B]
also make good sense. But the former makes sense qua conjunction of "I intend to do A" and "I intend to do B" in their intention ascribing roles (see paragraph 33 above). It by no means follows that the same is true when they are replaced by primary expressions of intention.

60. Another possible source of confusion is the fact that the punctuated text

My children shall have a good education.
I have already begun to put money aside,
can easily be confused with the conjunctive statement
My children shall have a good education and I have already begun to put money aside.
The latter would be a prima facie counter-example to the thesis I have been advancing.

61. Can it seriously be maintained that the only sound conjunctive expression in the neighborhood has the form

Shall be [p and q]?

62. To see why the answer is Yes, it is necessary to reflect on a feature of ordinary logic which is so familiar that it is treated with contempt: Conjunction Introduction (CI).

63. I called attention above to the distinction between a text and a conjunction. An example of a text would be

The weather was clear. The temperature was low.
Here there is punctuation but no logical connective, no "and." We have two sentences rather than one conjunctive sentence. CI tells us that
'p' and 'q' together imply 'p and q.'
It is a principle of implication which takes one from separate sentences to a single conjunctive sentence.

64. Thus the above text implies the conjunction

The weather was clear and the temperature was low.

65. In the opening paragraphs of this essay I commented on the connection of logical concepts and implications with propensities to believe and not to believe, and on the pressures they exert in the direction of coherence and the absence of inconsistency. In CI we find a related pressure, this time one which constrains us to bring together statements and beliefs which, though they coexist, are not confronting one another, and combine them into a functional unity in which they rub shoulders; into a single statement or belief whose implications can be drawn in accordance with logical principles.14

14 There are also the pressures and constraints of relevance, about which much would have to be said to get significantly above the intuitive level. If I engage in hand-waving in introducing a pressure of relevance, I am at least in good company.

66. Before discussing the bearing of CI on practical reasoning, on reasoning about purposes and values, one more principle which concerns an implication tied directly to "shall" (in my idiolect) remains to be introduced.

So-Be-It: "Shall be [φ]" and 'p' imply "Shall be [φ and p]" where 'φ' is a formula which may or may not be logically complex.
Implications conforming to this principle push in the direction of getting relevant beliefs into the scope of our purposes and intentions, and hence into the scope of practical reasoning.15

15 It is important not to confuse the concept of intention with that of desire or motive. A state of affairs can be a constituent of an intention without being an object of desire or a source of motivation. For an examination of key topics pertaining to the relations between the concepts desire, intention, and satisfaction, see my "Thought and Action" in Lehrer (ed.) Freedom and Determinism, New York 1966. See particularly Section III.

67. Notice that So-be-it requires that

Shall be [I will do A]
where "will" is the Standard "shall," not have the same sense as
Shall [I do A].
My doing A in the future can enter into an intention either as a target16 of an intention or, via So Be It, as something which I believe will happen, as when, confronted by a highwayman, I believe, fatalistically, that I will hand over my money. A similar distinction might be drawn between
Shall be [my lawn will be mowed]
and
Shall be [my lawn be mowed]
where the infinitive indicates that this content is not present in the intention by virtue of So-be-it. But these niceties would not contribute to the solution of the problems with which I am concerned.

16 Roughly speaking, the target of an intention is that which is being viewed sub specie "up to me." For an elaboration of this distinction see "Some Reflections on Contrary to Duty Imperatives," Nous I, 1967, pp. 315 ff. For its application to an argument of Castaneda's, see "Volitions Re-affirmed," Action Theory, ed. by Myles Brand and Douglas Walton (Dordrecht, 1976), p. 62. For a caveat on degrees of up-to-me-ness, see paragraph 82 below.

68. These considerations support the principle

Shall-Be: "Shall [I do A]" implies "Shall be [I do A]."
Note that this infinitive phrase "I do A" presents the action as up to one, and not, like "Shall be [I will do A], "sub specie So-Be-It. Thus, the converse of Shall-Be as formulated above is also true, which it would not be if the implicate were "Shall be [I will do A]." It is Shall-Be which, together with S-IMP underwrites the intuition expressed in paragraph 45 above.

69. It might be thought that the distinction between what I intend to do and what I believe I will do is "merely psychological." This is of a piece with claiming that in general the distinction between believing and intending is "merely psychological." It is the "merely," of course, which is out of place. Is the difference between expecting ("It will rain") and retrospecting ("It did rain") "merely psychological" -- or are there conceptual points to be made about tenses? Clearly the latter, and this is the parallel on which I have insisted.

70. As for CI, in practical contexts, it is simply a consequence of the above basic principle, where this has been generalized to cover the implications of two or more sentences, thus

If P1,' 'P2,'... 'Pn' together imply 'Q,' then "Shall be [P1]," "Shall be [P2],"... "Shall be [Pn]" together imply "Shall be [Q]"
In the case of CI we would have 'P1,' 'P2,'... 'Pn' together imply 'P1 and P2 and . . . Pn'
hence
"Shall be [P1]," "Shall be [P2],"... "Shall be [Pn]" together imply "Shall be [P1 and P2 ... and Pn]"

VII

71. The fact that So-be-it involves the introduction of beliefs into the contents of intentions calls for a review of our regimented formulation of intendibles.

72. One must distinguish between those constituents of a complex intendible which are there merely as, so to speak, factual connective tissue, i.e. items which one who adopts the intention believes to be the case, and those constituents the target of the intention, i.e. the presence of which makes it an intention.

73. One might think that the distinction in question coincides with that between items which are present as sources of motivation, either positive or negative, and items which are motivationally neutral. But that would be a mistake.

74. In the first place, a constituent of a scenario can be something that is believed to be the case and yet be directly relevant to a preference between that scenario and its alternatives, thus

Shall be [... Jones has insulted me and if I see him, I frown ...]
and
Shall be [... Jones has insulted me and if I see him, I smile ...]

75. And in the second place, one wants to allow for the possibility that one and the same intention can be carried out for different motives. This requires a way of representing intentions which abstract as far as possible from the distinction between motivating and non-motivating factors.

76. We must distinguish between questions concerning the nature and logic of intentions, and questions concerning the explanation of why a person adopts a certain intention or prefers one course of events to another.

77. Of course, the representation of an intention can be combined with a representation of motivational structure by the use of such words and phrases as "in order to," "in spite of" and "because" -- thus,

Shall be [I do A in order to bring about X, in spite of the fact that X would also bring about Y, because X would enable me to do B].

78. What we are looking for is a way of representing complex intentions which abstract from the structure introduced by such words and phrases.

79. From this point of view, the basic distinction between intention-factors is that between those which are viewed sub specie "up-to-me," and those which are viewed as given, i.e. those items which are there by virtue of So-be-it. And this distinction can be represented grammatically by using verbs in the infinitive for the former, and statement-making sentential forms for the latter, thus

Shall be [I do A]17
as contrasted with
Shall be [...and I will do A]
and
Shall be [my lawn be mowed]
as contrasted with
Shall be [... and my lawn will be mowed].

17 Notice that the entire conditional
If I do A, I will be punished
is a statement-making form, and hence is appropriately changed into an intention by So-be-it, thus
Shall be [... If I do A, I will be punished].

80. Notice that in each case the So-be-it item is introduced by "... and." This is because every intention must have at least one "up-to-me" constituent.

81. The "up-to-me"-ness of intendible constituents can be made explicit by prefixing them with the conditional "if it is up to me," thus

Shall be [if it is up to me, I do A]
Shall be [if it is up to me, my lawn be mowed].
This antecedent would get affirmed by virtue of So-be-it.

82. Problems concerning the criteria for determining what is up to an agent and, perhaps, to what degree or extent are notoriously difficult. Fortunately (like issues pertaining to probability) they do not bear on the central topics of this paper -- so I simply look them in the face and move on.

VIII

83. We are now off and running. With the principles already presented (along with constraints of relevance) we have a logical pressure to bring our separate intentions together with our separate beliefs into one picture or scenario. Thus consider the text,

I shall go downtown tomorrow, if the weather is clear. The weather will be clear. If the weather is clear, the streets will be crowded. If the streets are crowded, I will be miserable.
This text implies, by virtue of Shall-Be, So-be-it and CI, the conjunctive intention
Shall be [I go downtown tomorrow, if the weather is clear and the weather will be clear and if the weather is clear, the streets will be crowded and if the streets are crowded, I be miserable].

84. Notice that of the contents of this intention, "If the weather is clear tomorrow, I go downtown" is present as an intendible derived by CI from the conditional intention at the beginning of the text, whereas "if the weather is clear, the streets will be crowded" is a prepositional constituent via So-be-it. Notice also that while "if the streets are crowded, I will be miserable" is introduced as a whole via So-be-it, its consequent is represented as an intendible state of affairs as being "up to me."

85. The above conjunctive intention, in turn, implies, by three subordinate uses of Modus Ponens, the intention

Shall be [I go downtown tomorrow and be miserable].
I ask myself if I accept this intention. If I reject it, I am logically committed to rejecting either
Shall be [I go downtown tomorrow]
or
Shall be [I be miserable tomorrow].
Presumably I will reject the former.

86. Notice that I can also ask myself

Is it really true that if the streets are crowded and I am there, I will be miserable?
and decide, on reflection, No; in which case that conjunct drops out and I am left with
Shall be [I go downtown tomorrow if the weather is clear and the weather will be clear and if the weather is clear, the streets will be crowded]
which implies
Shall be [I go downtown tomorrow and the streets will be crowded]
which, in turn implies, by simplification,
Shall be [I go downtown tomorrow]
and, finally, since "shall be" is no longer serving to embrace So-be-it factors,
Shall [I go downtown tomorrow]
which, if I accept it, is the intention to go downtown tomorrow.

87. Notice that it might occur to me as relevant that when my friend Jones is with me I am happy, even though the circumstances are otherwise unpleasant. Thus, even if I retain my belief that if the streets are crowded and I am there I will be unhappy in the weaker form

If the streets are crowded, and I am there, then other things being as expected, I will be unhappy.
I now see that other things are not equal in that
If the streets are crowded and Jones is with me, I will not be unhappy.

88. At this point I may wish and, indeed, if the situation can be realized, intend that Jones be with me tomorrow. Furthermore, I may believe that Jones will be with me if and only if I invite him. So-be-it takes me to

Shall be [Jones be with me and Jones will be with me if and only if I invite him]
I am confronted by the choice
Shall be [I invite Jones and he be with me]?
Shall be [I not invite Jones and he not be with me]?
If I opt for the former, simplification takes me to
Shall be [I invite Jones]
and, hence, by Shall-be (converse), to
Shall [I invite Jones].

89. The fact that I intend to invite him implies, by virtue of the basic principle pertaining to "shall" (paragraph 43), that ceteris paribus I will in point of fact invite him. Believing that other things are in fact equal, I can be expected to believe that I will invite him, and, hence, to believe that if I go downtown, Jones will be with me. So-be-it inserts this into the intention matrix as follows,

Shall be [If the weather is clear, I go downtown tomorrow and the weather will be clear and if the weather is clear, the streets will be crowded and if I go downtown, Jones will be with me and if the streets are crowded and Jones is with me, I be happy]
which implies
Shall be [I go downtown tomorrow and be happy]
which in turn implies
Shall [I go downtown tomorrow].

90. Of course, if I learn that Jones will not be able to accompany me, the question

Shall [I go downtown tomorrow]?
will be reopened -- unless I am bored by the whole business.

91. The example we have been considering should remind us that an essential feature of practical reasoning is the fact that it constantly involves an appeal to preference.18 This occurs when one is confronted by alternatives, thus

Shall I do A?
Shall I do B?
e.g.
Shall I eat the yogurt?
Shall I eat the ice cream sundae?

18 For a more detailed account of preference, see "Thought and Action" in Keith Lehrer (ed.), Freedom and Determinism, especially pp. 113 ff. and pp. 131 ff. The role of CI was smuggled in under the name "the summative phrase." I there emphasize -- which I do not here -- that the summative phrase constructs alternatives each of which includes the implications of not choosing the others.

92. The following relevant considerations occur to me

The yogurt is plain. Plain yogurt is blah.
The yogurt has less calories.
The ice cream sundae is very fattening.
The ice cream sundae is very tasty.
CI and So-be-it expand the content of the above alternatives, thus
Shall be [I eat the plain blah yogurt with few calories]?
Shall be [I eat the tasty ice cream sundae which is very fattening]?

93. Confronted by these alternatives I may opt for one or the other, in which case that is my preference on the occasion. Or I may remain in a state of indecision, casting about for other alternatives or additional relevant considerations pertaining to the alternatives at hand. Thus it may occur to me that I have to work in the garden tomorrow, and so can stand a few more calories. The above alternative becomes

Shall be [I eat the plain blah yogurt with few calories which I shall more than use up in the garden tomorrow]?

Shall be [I eat the tasty ice cream sundae which, though very fattening in general, will not fatten me since I will use the calories up tomorrow working in the garden]?

The choice of the ice cream sundae is now humanly predictable. I opt for the latter alternative, which implies
Shall be [I eat the ice cream sundae]
and, hence,
Shall [I eat the ice cream sundae].

94. Thus CI together with So-be-it, takes our separate purposes and relevant beliefs and puts them together into encompassing alternatives:

1. Shall be [I do A at t, which means that . . .]?
2. Shall be [I do B at t, which means that . . .]?
3. Shall be [I do C at t, which means that . . .]?
The successive steps are
  1. elaboration by CI and So-be-it, and the drawing of implications
  2. choice, e.g. (2), or, continuing indecision
  3. simplification. Shall be [I do B at t]
  4. intention to act. Shall [I do B at t]
which, when time t comes (and I do not change my mind), generates a doing (or an attempt to do) B.

95. This picture is one according to which practical reasoning is essentially the process of elaborating alternative scenarios for a choice.

96. The elaboration is "rational." What about the choice? Is there any sense in which the choice can be "rational" over and above that in which it is so by virtue of being a choice between rationally elaborated alternatives, i.e. in which relevant considerations and relevant implications have been taken into account? Let me postpone this question until I have explored another line of thought.

IX

97. We have been considering implications between purposes or intentions

I1 implies I2
where I1 is, for example, "Shall be [I go downtown tomorrow, if the weather is clear, and the weather will be clear]" and I2, "Shall be [I go downtown tomorrow]." The implication tells us that the combination
Accepting I1 and not accepting I2 (or accepting the contradictory)
is unreasonable. Thus the combination
accepting "Shall be [I go downtown tomorrow, if it is clear and it will be clear]" and not accepting

"Shall be [I go downtown tomorrow]" (or accepting "Shall be [I not go downtown tomorrow]")

is subject to logical criticism.

98. But, of course, the fact that the implication tells us that this combination is unreasonable tells me nothing about the reasonableness or unreasonableness of accepting I1 or of accepting I2.

99. The implication, however, can be interpreted as telling us that accepting

"Shall [I go downtown tomorrow]"
is reasonable relative to accepting
"Shall be [I go downtown tomorrow, if it is clear and it will be clear]."
Relative reasonableness, however, is not what we were looking for when asking our postponed question. What we had in mind was rather a reasonableness which is not in the above way relative -- and which we might be tempted to characterize either as "absolute" or as "intrinsic."19 Can we find it? Let us follow Hume's example and beat about in the neighboring fields.

19 These two tags are, of course, provisional, for until we clarify the concept of relative reasonableness, it remains unclear what it is to be contrasted with.

100. What is the relation between

I shall do A
and
I ought to do A (or I should do A)?
The answer to this question must surely run somewhat as follows: The latter makes a positive assessment of the intention expressed by the former.

101. Since to make a "positive assessment" of something amounts to characterizing it as in some sense good, and since there is clearly some sort of conceptual tie between "good" and "ought" (or "should"), as illustrated by the truistic feel of "one ought (or should) choose the good," the first fruit of our beating would seem to be that if we can get a grip on "ought" or "should," we might get a grip on "good," and, perhaps, on our "absolute" or "intrinsic" reasonableness as well.

102. Consider the statement

If Jones wants X, he should (ought to) do A
Clearly the underlying theme is that given Jones's circumstances, C, and a proper spelling out of just what it is that Jones wants,
"Jones gets X" implies "Jones does A"
or, equivalently,
"Jones's getting X" implies "Jones's doing A, if Jones is in C"
because, for example, a person's doing A is causally necessary for their getting X, if they are in C.

103. Suppose that Jones does want X, i.e. his purpose or intention is to get X; can we reason

If Jones wants X, he should (ought to) do A, since he is in C
Jones wants X
So, Jones should (or ought to) do A?
If so, we would have arrived at a positive evaluation of Jones's doing A from a factual premise (Jones "wants X") together with a premise the gist of which is a causal connection between an outcome and an action. In other words we would ostensibly have derived a value from matters of fact, an ought from ises.

104. But, of course, this is an illusion. We have simply run into another case of relative reasonableness. The hypothetical proposition

If Jones wants X, he ought to do A, if in C
which is based on, but not identical with, the causal implication in question, tells us that
"Shall [I, Jones, get X]" implies "Shall [I, Jones, do A if in C]"
i.e. that the intention
"Shall [I, Jones, do A, if in C]"
is reasonable relative to the intention
"Shall [I, Jones, get X]"
by virtue of the implication
a person's getting X implies their doing A if in C.

105. Thus, given that Jones wants X, i.e. intends "Shall [I get X]" all we are entitled to conclude is that

"Shall [I, Jones, do A]"
is implied by and hence reasonable relative to an intention which Jones actually has. But, again, this is not what we were looking for -- though it may be all we can get. Nothing is implied about a "non-relative" reasonableness of getting X. In spite of the "oughts" and "shoulds" there is not the ghost of a hint as to how the intention to get X might be "non-relatively" reasonable.

106. Some light is, indeed, thrown on the goodness of objects and instrumentalities by the concept of relative reasonableness. Thus part of the conceptual tie between "good" and "should be chosen" is captured by so-called "hypothetical imperatives, "i.e. statements of the form we have been considering. Clearly there is an intuitive connection between "juicy apples are good" and "if a person wants enjoyable apples, they should choose juicy ones."

107. Yet, after all, we distinguish "goods" from "necessities (relative to a purpose)." Thus, strictly speaking, the core of

If Jones want S, he ought to do A
is
If Jones wants X he must do A
for that is the proper force of the implication on which it is grounded, which tells us that Jones doing A is a necessary condition of Jones getting X.

108. And the same is true of

If Jones wants an X which is φ, he should choose one which is ψ.
The "must" can be weakened by a ceteris paribus clause to allow for optional alternatives, but this root idea remains that of a necessary condition.

109. And, after all, there is an air of paradox about

If Jones wants to shoot people quietly, he should (ought to) use a silencer.
We would feel more comfortable with
If Jones wants to shoot people quietly, he must (would have to) use a silencer.

110. Reflection on this fact suggests that when either "ought" or "should" is used in the phrasing, the implication is that the "want" clause refers to a purpose which is in some way connected to an intrinsic good.

111. Thus we are willing to say that

If Jones wants to be healthy he should (ought to) eat a balanced diet
where we think of eating a balanced diet as a good by virtue of its relation to health. More, however, seems to be involved than that the intention to eat a balanced diet is reasonable relative to the intention to be healthy, for the goodness of eating a balanced diet seems to be grounded in the goodness of health. Is health, in turn a good by virtue of its relation to something else? Is not our search for a non-relative reasonableness equivalent to a search for an intrinsic good?

112. And, indeed, we do think of instrumentalities as good because we think of the intention to have them available as reasonable relative to an intrinsically good purpose. But what might such a purpose be?

113. And is the intrinsic goodness of a purpose (i.e. a purposed -- intended -- state of affairs) a matter of a non-relative reasonableness of the purposing -- intending?

X

114. Consider the intention expressed by

Would that I lived a satisfying life all things considered.
It has long been argued20 that what pattern of life would satisfy a given person is a question of fact -- difficult to be sure, but in principle capable of being given an objective or, in a non-pejorative sense, scientific answer.

20 Since the time of the Platonic Socrates. For a more detailed analysis of the issues about to be raised, see "Reason and the Art of Living in Plato," in Dale Riepe (ed.), Phenomenology and Natural Existence: Essays in Honor of Marvin Farber (Albany), 1973 [reprinted as Chapter 1 in Essays in Philosophy and its History] and "On Knowing the Better and Doing the Worse," International Philosophical Quarterly, vol. 10, 1970 [reprinted as Chapter 2 in EPH], and also Chapter 7 of Science and Metaphysics.

115. It has also been argued, for an equally long time, that of necessity a person does seek his own happiness. The oddness of the intention

Would that I did not lead a satisfying life, all things considered
has been construed as a matter of a psychological impossibility of having this intention.

116. Indeed there has been a tendency to think that as a matter of psychological necessity, all deliberate or reflective (as contrasted with impulsive) action is grounded in an intention of which

Would that I led a satisfying life, all things considered
is the over-riding constituent in the sense -- explicated in terms of the distinctions I have been drawing -- that in particular circumstances the rationally elaborated alternatives in which the idea of leading a satisfying life, all things considered, occurs, are very schematically,
(α) Shall be [I lead a satisfying life, all things considered, which means p and not-q and ... and that I do A and that I do not do B]?

(β) Shall be [I do B which means that I not do A and q and not-p and ... and that I not lead a satisfying life all things considered]

of which only the former can, as a matter of psychological necessity, be accepted.

117. Of the mistakes which spoil the insight contained in this tradition, the most obvious is that which posits that the intention

Shall be [I lead a satisfying life, all things considered]
is a constituent in all rationally elaborated alternatives (i.e. in all non-impulsive action).

118. Notice that the phrase "satisfying life all things considered" represents what the scholastics called a formal end. As such it has only such force in particular practical reasonings as a "namely rider" would provide. Thus the text of the above intention with respect to a particular occasion should read, schematically,

Shall be [I lead a satisfying life, all things -- namely: p, q, r, . . . -- considered].

119. Yet there is a more basic mistake, as is shown by the appeal to psychological necessity. The reader will remember my caveat, in paragraph 69 above, against supposing that the difference between intending and believing is "merely psychological." A related point needs to be made here.

120. Is the oddness of accepting alternative (β) above a "mere" matter of psychological impossibility? Is one simply appealing to the principle "ought implies can" in the form "what can't be, can't be reasonable" to conclude that alternative (β) can't be reasonable? After all, it might be said, an intention can't be reasonable unless it can be had!

121. But, of course, it wouldn't follow from the non-reasonableness of (β) that (α) was reasonable.

122. And, indeed, in what sense could (α) be a reasonable intention? Perhaps because of the phrase "all things considered." For we have seen that the principle of taking all relevant facts and purposes into consideration is built into the logic of practical reasoning.

123. Thus it would seem that intentions which are arrived at by taking all relevant considerations into account are, in so far forth, reasonable.21

21 I leave to other occasions the problem of mobilizing such distinctions as those between known relevant considerations, available relevant considerations, potentially available relevant considerations, and between objective reasonableness, subjective reasonableness, etc. The point is to get reasonableness into the picture.

124. If, however, the formal intention

Shall [I promote my own happiness, all things considered]
were intrinsically reasonable, then the intentions which are implied by it would also be reasonable, period. This is to say, they would not merely be reasonable on the hypothesis that something else is reasonable, but reasonable because that something else is reasonable. We must distinguish between categorical reasonableness (of which intrinsic reasonableness is the prime mover) and relative (or hypothetical) reasonableness. Reasonableness (like theoremhood) can be simultaneously categorical and derivative.

125. And if all this is correct, then the idea that the intention that one lead a satisfying life is a psychologically necessary intention can be traced to a misconstrual of the connection, pointed out in the opening paragraphs of this essay, between "seeing" implications and having ceteris paribus propensities to exhibit uniformities in one's modes of thought.

126. Thus it is not psychologically impossible to opt for alternative (β). All we need say is that the "pressure of reason" is against it. And the source of this pressure lies not in the formal end "satisfying life all things considered," but in the conceptual pressure to include in the elaboration of this intention on this occasion all relevant specific considerations, i.e. to enrich the "namely" rider.

127. The "specific" considerations deemed relevant on any particular occasion will inevitably be, to a greater or lesser extent, fragmentary and generic. But this humanly unavoidable fact would be a flaw only in the practical reasoning of an angel.

XI

128. That it is not psychologically impossible to fly in the face of intentions having the formal structure "would that I led a satisfying life all things considered," is important, not only because it is true, but because it opens the door for a class of intentions which have logical features not yet considered, and which provide the framework for an understanding of the moral point of view.

129. And, indeed, if one places the conceptual considerations advanced in the preceding section in the context of classical moral philosophy, it is immediately apparent that they constitute a sophisticated form of what has come to be called Rational Egoism according to which the fundamental intrinsic value is one's own happiness, all things considered.

130. Rational Egoism, of course -- except in its most extreme form -- acknowledges the existence of sympathy and benevolence, and even grants that sympathetic and benevolent purposes can be wide in scope and intense in feeling. Yet it denies that they can generate a mode of reasonableness which is other than the relative reasonableness of such hypothetical imperatives as

If one wants to help widows and orphans, one should, for example, support laws against fraud and embezzlement.

Intrinsic reasonableness, the Egoist argues, requires an over-arching purpose which can find a place for both self-oriented and other-oriented "particular" purposes -- and this, it is argued, can only be provided by the pursuit of one's own happiness, all things considered.

131. Yet we saw that the "can" is suspect. Is it a matter of psychological possibility? That is, is Self Love the only over-arching purpose one can have? Or is Self Love the only over-arching purpose that can generate an intrinsic reasonableness? One suspects that by granting the existence of "particular" sympathetic and benevolent intentions, the Egoist has weakened the traditional argument from the supposed psychological impossibility of "cool hour" altruism. What, then, about reasonableness?

132. That informed self-interest and the moral point of view do, on occasion, conflict is a fact which is built into the human situation. To understand the nature and possibility of this conflict, we must take into account a further dimension of practical reasonableness -- one which, indeed, is closely tied to the very nature of reasonableness itself.

133. The challenge we face is, in essence, the following: Can we preserve the general strategy of the argument to date, and yet take distinctively moral values into account? I think so, and I will begin to sketch a line of thought which, I believe, adds the necessary dimension to the logic of purposes and intentions.

134. Consider the wish,

Would that I were happy.
This wish is obviously I-centered. How exactly is this I-centeredness to be understood? What feature of the expression of the wish embodies it?

135. To begin with, there is the obvious fact that the subject matter of the wish is formulated by means of the context dependent referring expression "I." The belief expressed by

I am in Houston
is in the same way I-centered.

136. If Jones also says

I am in Houston.
there is a sense in which he is saying the "same thing" I did and a sense in which he is saying something different. As Strawson would put it, he is using the same sentence to make a different statement. To this we could add that he is ascribing the same property, being now in Houston, to himself that I ascribed to myself.

137. Let us put this by saying that the wish and the belief are referentially I-centered.

138. Thus, if King Jones thinks

Shall be [I possess Paris]
and King Smith thinks
Shall be [I possess Paris]

they have, in a sense, the same intention, and, in a sense, different intentions. This fact makes possible a familiar witticism. Each can truthfully say: that which he wants, I want also.

139. From this perspective, the intention expressed by

Shall be [the nations remain at peace]
would not be referentially I-centered.22 Yet if the argument of the early sections of this essay is correct, there is a way in which this expression of intention is covertly subject matter I-centered. For we found it to be a conceptual truth that, given that its being the case that-p implies my doing A,
It shall be the case that-p
implies, ceteris paribus,
I shall do A.

22 Of course the expression is implicitly I-centered in that it directly or indirectly involves the use of indicator words ("indexicals"). But tokens of spatial indexicals (e.g. "here") and temporal indexicals (e.g. "now") can have the same reference when used by different persons, whereas, truistically, "I" cannot. And the important way in which expressions of intention can be covertly I-centered, is tied to this latter fact.

140. In other words, even though the manifest content of 'p' is not referentially I-centered, expressions of intention of the form

Shall be [p]
have, by virtue of the conceptual grammar of "shall," a conceptual tie to expressions of intention of the form
Shall [I do A]
which latter is referentially I-centered.

141. After all, it is a conceptual point about the universe of intentions that its point d'appui lies in intentions to do and, in the last analysis in the I-here-now intentions to do which are volitions, and, paralysis aside, culminate in action.

142. We might put this by saying that all intentions are explicitly or implicitly action I-referential.

143. To put it in this way, however, obscures the fact that although "in the last analysis" all action is action by individuals, and although all volition has "in the last analysis" the form

Shall [I do A here and now]
there is an important sense in which even such intentions can transcend I-centeredness.

144. The key point lies in the fact that one can think of oneself as a member of a group. This obvious truth is tied to certain features of the conceptual grammar of reference to groups.

145. I can refer to myself either verbally or in foro interno as Wilfrid Sellars, as the author of "Empiricism and the Philosophy of Mind," or as I.

146. Similarly I can refer to a group as the Society for the Prevention of Cruelty to Animals, as the organization which won the Pittsburgh Civic Medal last year, or I can refer to it as we. Which group is, on any occasion, referred to as we can be made explicit by a modifier, thus we Texans, we members of the American Philosophical Association, etc.

147. But before we can put these familiar points to good use, they must first be elaborated. We must take into account certain other features of the logical grammar of "we."

148. In the first place, there is the obvious fact that we includes I.

149. Then there is the related fact that actions can be ascribed to individuals not only where they are picked out by singular terms, e.g.

Socrates ran. The president ran
but also where they are referred to by using the apparatus of quantification, thus
Everybody ran
or, more interestingly,
Each of us ran.

150. Consider, therefore, the intention

Shall be [each of us will run a hard race].
Notice, in the first place, that this intention is intersubjective (with respect to us) in the sense that any of us can have this intention, where, qua state of affairs intended it is, in a stong sense, the same. In this respect it resembles
Shall be [the nations remain at peace]

151. But it has already been suggested that the latter is action I-centered. Thus, consider the dialogue

Jones: Shall be [the nations remain at peace]
Smith: Shall be [the nations remain at peace].
Although Jones and Smith are expressing a shareable state of affairs intention, the former, as expressed by Jones, implies, given certain relevant considerations,
Shall [I, Jones, do A]
and the latter
Shall [I, Smith, do B].23

23 Remember that what is implied by Jones's (or Smith's) intentio to have the implied intention.

152. We can put this by saying that a shareable state of affairs intention (where this amounts to more than a shareable wish) implies, for each person by whom it can be shared, referentially I-centered action intentions.

153. Consider, next

(α) Shall be [each of us have a new car].
This subject-matter shareable (among us) intention obvously implies every intention the expression of which results from replacing "each of us" by a designation of a member of us, e.g. Tom, thus
(β) Shall be [Tom have a new car].
Thus, each person who has shareable intention (α) has an intention which implies intention (β), which is also a subject-matter shareable intention.

154. On the other hand, each person (e.g. Dick) who has intention (α) has an intention which implies for him an intention of the form, (γ)

(γ) Shall be [I, Dick, have a new car]
which, of course, is not a subject matter shareable intention.

155. We might be tempted to call it a referentially I-centered intention; and, indeed, to do so would be in accordance with our usage up to now. Yet if we consider it in relation to (α) which implies it, the phrase "I-centered" strikes us as inappropriate. For in practical contexts, the phrase "ego-centered" has an implication with respect to motivation which is inconsistent with the idea that Dick has reasoned

Shall be [each of us have a new car]
So, Shall be [I, Dick, will have a new car].

156. In his reasoning, so to speak, Dick has come into the picture as "one of us." We might, therefore, reformulate his intention as

Shall be [I, Dick, because I am one of us, have a new car].
Notice that here we are taking into account features of the content of intentions from which we have hitherto made abstraction. (See paragraphs 77-8.) It is important to realize that because is just as much a part of the content of the above intention as it is a part of the belief content expressed by
The Pirates should win because they are the better team
and has the same rationale.

157. Notice, next, that although (α) is a state-of-affairs shareable intention, it does not seem to be an action shareable intention, for as far as anything we have brought into the picture so far is concerned, its connection with action is via the fact that as a state of affairs intention it implies for each person who can share it, e.g. Dick, given that Dick's promoting the state of affairs each member of the group having a new car implies Dick's doing A, e.g. contributing to a fund,

(δ) Shall [I, Dick, contribute to the car fund]24

158. We can put this by saying that an intention can be we-referential qua state of affairs intention, and yet be action I-centered.

159. This would be a special case of the idea that an intention can be shareable in its manifest content, thus

Shall be [the nations remain at peace],
and yet be action I-centered in the sense that all its implications with respect to intentions to do are of the form
Shall [I, S, do A]

XII

160. We have seen (paragraph 141-2 above) that in a sense all intentions are explicitly or implicitly action I-referential. We are now in a position to ask the framework question: Is there any way in which an intention can be not only we-referential, but action we-referential?

161. To answer this question, we must retrace our steps and remember that way back before the concept of we-intentions was introduced, it was noted that a person can envisage one of their actions sub specie So-be-it, as when, for example, I represent myself as handing over money to a highwayman, thus

Shall be [... and I will hand over the money]
So-be-it
which contrasts with the sub specie "up-to-me" intention
Shall [I keep my cool].

162. A scenario which contains the latter is explicitly action I-referential, for it contains my doing A as an intention to do rather than as (via So-be-it) a factual constituent of an intention.

163. Thus consider the difference between

Shall be [... and most of us will not resist temptation]
So-be-it
and
Shall [each of us resist temptation].
In the former, the action of resisting temptation enters into the intention as a constituent of a belief rather than as something intended.

164. Whereas in the latter, resisting comes in as something to be done. Thus the latter is surely the we-intention counterpart of the explicitly I-referential action intention

Shall [I keep my cool].

165. Schematically, the difference is that between

Shall be [... and each of us will do A, if in C]
So-be-it
and
Shall [each of us do A, if in C].

166. The following is a perfectly coherent intention Shall be [each of us do A and most of us will not do A].25

25 The reader should reflect on the intention
Shall [each of us not take the short cut this week unless more than ten of us will take it]
See my comment on the "generalization argument" in Science and Metaphysics, Chap. VIII, sec. xiii.

167. It is essential to note the the intention

Shall [any of us do A, if in C]
must not be confused with the intention
Shall [everyone of us do A, if in C].
The latter is an intention which is realized only if everyone of us who is in C does A. The former is realized even if only a single one of us who is in C does A.26

26 Thus, for
Shall [any of us do A, if in C]
to be a proper intendable, it need not be practically possible that everybody does A, if in C.

168. It must also be carefully borne in mind that although the concept of a group intention and a group action is a perfectly legitimate one, the action we-referential intentions we are considering are intentions had by individuals. It is individuals who intend

Shall [each of us do A, if in C]
.

169. We have seen that the intendible constituents of an intention (as contrasted with the So-be-it constituents) are there sub specie "up to me." One might therefore be tempted to expostulate: "What others do is not up to me. How, then, can 'Shall [each of us do A]' express a proper intention, since 'each' refers to others than oneself?"

170. A superficial answer would point out that on occasion what another does is, at least to some extent up to me. That we can influence people is as relevant to practical reasoning as that we can influence sticks and stones.

171. The answer which meets the question, however, involves the fact (noted in paragraph 81) that to say that the intendible consituents of an intention are those which are present sub specie "up to me" is equivalent to accompanying them with the conditional "if it is up to me," thus

Shall be [if it is up to me, I do A].

172. Now "up to me" is the first person form of "up to X." Obviously other referring expressions can replace "X." Thus the correct answer to the above challenge consists in calling attention to the fact that the "up to the agent"ness of action we-referential intention is to be formulated as follows,

Shall [each of us, if it is up to them, do A]
and this in no way requires that what others do be up to me.

173. Notice, finally, that the circumstances in which each person is to act may include (via So-be-it) a reference to what it is believed that others will in point of fact do, thus

Shall [each of us pick up the torch, if the one of us to the right of them drops it]
taken together with "My neighbor to the right has dropped the torch" implies, via Shall-be and So-be-it,
Shall be [each of us pick up the torch if the one of us to the right of them drops it and my neighbor to the right has dropped it]
and hence, via the instantiation move from "each of us" to "I," Modus Ponens and Shall-be,
Shall [I pick up the torch].

XIII

174. We are almost through with the necessary hair-splitting. As a matter of fact all that remains to be done is to combine the above distinctions with those which were drawn in the first part of this essay and apply them to some intuitively relevant cases.

175. Thus, consider the intention

Shall be [whooping cranes survive].
This is a shareable intention available for members of the Whooping Crane Society, as is
Shall be [each of us WCS members do what he can, ceteris paribus, to promote the survival of whooping cranes].

176. And consider the general hypothetical imperative,

If a member of the WCS wants to promote the survival of the whooping crane, he should pay his dues and obey the by-laws.
This tells us that
"Shall [each of us members of the WCS pay his dues and obey the by-laws]"
is reasonable relative to
"Shall [each of us members of the WCS promote the survival of whooping cranes]."

177. We have an implication which relates one we-referential action intention to another.

178. We-referential action intentions imply intentions to act on the part of individuals, not simply by virtue of the general principle which relates "Shall be" to "Shall I do," but directly by virtue of the relation between "Shall anybody do" and "Shall I do."

179. If the reasonableness of an available intention of the form

Shall [I do A]
is a matter of its reasonableness relative to (i.e. its being implied by) an intention of the form
Shall [any of us do ...]
this might be indicated by flagging the "shall" of the former intention with a subscript "we," thus
Shallwe[I do A].
Thus understood, the later is equivalent to the schema
Shall [I do A, because shall any of us do ...]
At this point paragraph 156 should be reviewed.

180. Correspondingly, if a person has the intention

Shall [I do A]
and has it because it is implied by the intentions
Shall [any of us do A]
which he also has, we can say that Jones intends to do A sub specie "one of us," and flag our representation of his intention with a subscript "we," thus,
Jones intends "Shallwe [I do A]"

181. If I have the intention

Shall be [whooping cranes survive]
I may infer, first
Shall be [the Whooping Crane Society survive]
and, then, by virtue of the connection between state of affairs intentions and intentions to do (together with relevant considerations)
Shall [I pay my dues].
In this case, my intention, though it concerns the WCS, of which I am a member, is not from the point of view of a member of the WCS.27

27 Notice that the WCS can be said to have a point of view as an organization, just as it can be said to act as an organization. The concept of groups purposes, intentions and actions is an important and highly relevant one which, however, I am skirting on the present occasion.

182. Thus which of the implication structures I 'pick up' in my deliberations about paying my dues determines whether the point of view from which I am reasoning is "private" or that of a member of the group.

183. Notice finally, that a shareable state-of-affairs intention will, in general, have implications with respect to intentions to do of both the I-referential and the we-referential kind. (Since the latter, in turn, imply I-referential action intention, we might distinguish we-derivative from primary I-referential action intentions.) The facts by virtue of which these implications obtain may well overlap, though they cannot coincide.

XIV

184. But the facts that we-referential intentions can be shared and that an action intention can be we-referential are logical points which establish at best a necessary condition for the possibility of a distinctively moral point of view. What can be added to turn this necessary condition into a sufficient condition?

185. The dialectics of our elaboration of Rational Egoism suggests that what we are looking for is a non-relative reasonableness pertaining to we-referential action intentions. Is there such a thing?

186. That there is, is suggested by the fact that the purpose or intention expressed by

Shall [each of us WCS members promote the survival of whooping cranes]
is reasonable for members of the WCS, because it is by virtue of that shareable intention that there is such a thing as the WCS.

187. This intention defines what might be called "the WCS member's point of view." It is the prime mover intention of this point of view,28 and generates, in matters of factual implications and premises, subordinate intentions, the relative reasonableness of which is grounded in its intrinsic and non-relative reasonableness as constituting the WCS point of view.

28 Notice that there is a sense in which this intention can be construed as grounded in the intention
Would that whooping cranes survive!
which is a shareable intention, one, however, which abstracts from the distinction between we-referential and non-we-referential action intentions.

188. Of course, any member of the WCS can, and indeed often will, work out in his practical reasoning not only the implications of the WCS point of view for people who are in their particular circumstances, but also the implications of these circumstances for the personal point of view, or, in more traditional terminology, that of Self Love, i.e.

Shall [I promote my happiness].

189. Is there, then, an intention which defines the moral point of view? Is there an intrinsically reasonable we-referential action intention which stands to a moral community as the intention that "we" promote the survival of whooping cranes stands to members of the Whooping Crane Society?

190. Notice that while the intention that we members of the WCS promote the survival of whooping cranes is intrinsically reasonable for members of the WCS, the bare intention that whooping cranes survive, if reasonable, does not have this same kind of reasonableness. The point becomes obvious if we change the example to that of the Society for the Preservation of Architectural Landmarks.

191. In discussing Rational Egoism I distinguished between the formal intention

Shall [I do that which promotes my happiness, all relevant things considered]
and the specific contentual intentions which, in particular circumstances, "spell out" this "regulative principle" with namely riders.

192. I pointed out, in paragraph 124 above, that if the above formal intention is intrinsically reasonable, then the intentions which are implied by it are also reasonable, period. That is to say, they are not merely reasonable on the hypothesis that something else is reasonable, but reasonable because that something else is reasonable. Thus we distinguish between categorical reasonableness (of which intrinsic reasonableness is the prime mover) and relative (or hypothetical) reasonableness. Reasonableness, we saw, can be both categorical and derivative.

193. Our question thus becomes, Is there an intrinsically (and hence categorically) reasonable formal intention which is the regulative principle of the moral point of view?

194. Analogy would suggest the following:

Shall [each of us do that which, in the circumstances, promotes the happiness of each and every one of us, all relevant things considered].

195. ... Considering has as its state of affairs intention core

Shall be [each and every one of us leads a satisfying life, all relevant things considered]
which is the formal specification of an intrinsic and encompassing good.

196. Shareable by whom? Who are we? Though these questions are, in a sense, the same, let me begin with the second. Who are we? Not, it should be clear, the members of a corporate group, a group having corporate intentions and doing corporate deeds. The conceptual framework with which we are concerned is presupposed rather than constituted by "shoulds" and "oughts" which pertain to groups considered collectively.

197. This calls attention to the fact that a key feature of the above formal intention is the absence of an explicit membership rider for the "us" -- as contrasted with the "us members of the WCS" of the shareable intention we have just been exploring, and the "us Texans" mentioned earlier.

198. Is there, perhaps, an "implicit" and indeterminate membership rider, one which is left to be specified by the context on particular occasions?

199. Notice that what a person who is deliberating in the framework of this formal intention thinks their circumstances are is (obviously) not specified by the intention. The reasonableness on particular occasions of thinking that the circumstances are such and such is not to be equated with their being in fact thought to be such and such.

200. Can we similarly distinguish between what a person who is deliberating in the framework of this formal intention thinks of as us, and what it is reasonable for the deliberator to think of as us, all relevant things considered?

201. Of course, the formal intention does not specify what all these relevant things are, any more than the principle of Self Love tells one in particular circumstances the actual cash value of promoting one's happiness on the whole, all relevant things considered, in just these circumstances.

202. Thus, the absence of an explicit membership rider for the "us" does not mean that there is an "implicit" namely rider which is filled, in any given context, by a reference to whatever group a deliberator happens to "identify with" or think of as us (e.g. we WASPs).

203. The mere fact that a person identifies with WASPS no more entails that it is reasonable for him to identify with WASPs, than the fact that one believes oneself to be on terra firma entails that it is reasonable for one to believe this.

204. It is in the logical pressures generated by the principles which govern practical reasoning, and which find their expression in "all relevant things considered" that we find the point of the distinction between the moral community and those with whom we identify.

205. It will be remembered that I asked not only "who are we?" but "shareable by whom?" Do the above considerations throw any light on the latter question?

206. Surely once we have distinguished between the moral community and those with whom one who deliberates in the moral point of view happens to identify, we cannot stop short of identifying the moral community with that group with which it is reasonable for us, all relevant things considered, to identify. And, in the context of the formal intention

Shall [each of us do that which in the circumstances promotes the well-being of each and every one of us, all relevant things considered].
Who can that be, but rational beings generally?

207. Notice that I am not saying that everybody shares this shareable intention. I am simply saying that it defines the moral point of view, as contrasted with, say the WASP point of view. The intention

Shall [each of us WASPs do that which, in the circumstances promotes our common good, all relevant things considered]
is a shareable (by WASPs) we-referential action intention which is logically respectable. It simply does not constitute the moral point of view.

208. The question is it egoistically reasonable to acquire or maintain a propensity to intend sub specie "one of us" is a perfectly meaningful one. (Why should I be moral?) The fact that it is so in no way brings into jeopardy the autonomy of the moral point of view.

209. Finally the reader should bear in mind that this essay has been concerned with what a Kantian would call the formal structure of the moral point of view. In other words, I have been concerned with such issues as the intersubjectivity, urriversalizability, and reasonableness of categorical imperatives. For an exploration of how a utilitarian content fits into this Kantian form, the reader should consult the concluding sections (pp. 218-29) of Science and Metaphysics.29

29 London and New York, 1968.

Note: This paper was initially presented as the first of three Tsanoff lectures at Rice University in September of 1978. It was expanded and revised in response to conceptual pressures from Kurt Baier, Gerald Massey, and Don Morrison. I am grateful for the care with which they read and commented on the text. As a result a number of mistakes have been corrected and some topics brought to a sharper focus. That the main line of argument remains essentially unchanged may, of course, be due to pig-headedness.

University of Pittsburgh
Received January 2, 1979