Parfit, Reasons and Persons




Parfit is a ulilitarian.


The book is entitled Reasons and Persons because Parfit holds that "in various ways our reasons for acting should become more impersonal." (443)


This is argued in three ways:

1.     Moral action should aim to benefit the aggregate of people rather than specific people.  For example, actions which benefit specific people but harm all imperceptibly (pollution, overpopulation) may be wrong; also Parfit thinks we should reject principles favoring those emotionally closest to us.

2.     Since we may not be the same person through our entire lives, the notion of rational self-interest is flawed; our efforts to benefit should be aimed less at ourselves and more at people in general.

3.     Policies which aim to benefit specific people in the distant future may be self-defeating in that they ensure that such people will not exist.


It is the third claim that interests us.


Moral thinking about the distant future differs from traditional moral thinking in that it nearly always concerns decisions which affect the identity of the people they concern.




Chapter begins with the question, “what weight should we give to the interests of future people?”


On some views the interests of future people have little or no weight:

On some contractarian views only the moral interests of people who can reciprocate have any weight.

On some cost-benefit views the discount rate gives less weight to the interests of future people than of present people, which disvalues human life in the distant future almost entirely.

Parfit assumes these views should be rejected and that future people count as much as present ones do.   Temporal position has no influence on the moral considerability of one’s interests.


Does anyone have a problem with that?


The Non-Identity Problem—the problem of developing a moral theory (Theory X) that yields correct choices even when the choices affect who will exist.


Three kinds of choice:

1        Same-people choices:  the same people will have ever existed regardless of which action we take (most choices as ordinarily conceived, including life-and death choices)

2        Same-number choices:  different people will have existed if we take one action rather than others, but their numbers will have been the same (e.g., decision to have a child now or later) 

3        Different-number choices:  different numbers of different people will have existed depending on our choice (decisions affecting total population numbers).


Traditional moral thinking generally concerns same-people choices, but moral thinking about future generations usually concerns different-number choices. 


Same-number choices are an easier intermediate case. 


Chapter 16 develops the part of Theory X that deals with same-number choices.


The thesis of Chapter 16 is Parfit’s:

Same Number Quality Claim (Q):   In a same-number choice, if one of two options would result in a generally lower quality of life than would the other, it is worse—regardless of whether it makes any particular individuals better or worse off. (360)


The “regardless” clause embodies what Parfit calls The No-Difference View.


The main challenge to the Same Number Quality Claim is:

The Person-Affecting View (V):  If one of two choices lowers the quality of life for some particular people and the other does not make any particular people worse off, the first is worse.


We can grasp the difference between the No-Difference View and the Person-Affecting View via:


The example of two medical programs:  Two proposed medical programs have identical costs and effects, except that one would cure 1000 already existing fetuses of a handicap, while the other would instead of curing these fetuses prevent the same handicap in 1000 people yet to be conceived—people who would then owe their existence to the program. 


The Person-Affecting View implies that the policy which would prevent the handicap is worse, but the No-Difference View implies that these policies are morally equivalent.


In same-people choices, these Q and V coincide; they make a difference only if different people exist as a result of our choices.


Parfit’s argument against the person-affecting view is based on the:

Conservation/Depletion Example

     Now                 200 years



This is stipulated to be a same-number choice.


People's total quality of life is represented by the area under curves.


Q implies that conservation is the correct choice

          Total quality of life is better under conservation.  Whether individuals

          are better or worse off is irrelevant


V implies that depletion is the correct choice

Depletion is better for those who live in the first 200 years and worse for no one. 


Moreover if we assume that causing to exist is a benefit, depletion actually benefits everyone who ever lives, while conservation is worse for those who live in the near future.

This Parfit takes as a reductio of V.


Parfit does not seem to recognize that Q and V are not exhaustive.  We could have a view on which person-affecting considerations counted but were not decisive.


We are still looking for theory X.  Theory X would:

1.     solve the Non-Identity Problem in Different Number Choices

2.     justify Q




Parfit now considers and rejects seven candidates for Theory X.


Candidate theories may classified according to quantity or quality of value (let’s use the classical term “utility” here):

quantity = the sum total of utility

quality = average individual utility

Quality is quantity per capita


Seven Candidates for Theory X (see p. 403)

(Names in italics are Parfit’s; these are ordered in increasing importance of quality and decreasing importance of quantity)


Candidate View




1 Maximize total Utility (Impersonal Total Principle)


Is the only form of value

Has no value

Implies Repugnant Conclusion

2 Quality counts but may always be swamped by quantity

Has value

Has value, but a loss of quality can always be outweighed by gain in quantity

More plausible than 1, but still implies Repugnant Conclusion

3 Appeal to the Valueless Level

Has value, but only in lives  whose quality is above a certain level (though some lives below that level are worth living)

Has value; loss of quality cannot always be outweighed by gain in quantity

Implies (A) and (R)—variants of the absurd conclusion and the repugnant conclusions and mere addition paradox

4 Lexical View (cf. superman and herd)

There is no limit to the positive value of quantity. But in lives whose quality is mediocre, no quantity could be as good one life whose quality is blissful

Has value; loss of quality cannot always be outweighed by gain in quantity

Implies (A) and (R)—variants of the absurd conclusion and the repugnant conclusion and mere addition paradox

5 Diminishing Returns

As quantity increases, the value of extra quantity asymptotically approaches zero; but the limit at which this occurs is higher the higher the quality

Has value; loss of quality cannot always be outweighed by gain in  quantity

More plausible than 6 or 7; but refuted by the Absurd Conclusion

6 Limited Quantity View:   it will be worse if there is a smaller total of happiness than there might have been, unless this smaller sum is above a certain limit

Value of  quantity has an upper limit, which has been or can be exceeded

Has value; loss of quality cannot always be outweighed by gain in  quantity

More plausible than 7; but refuted by the Absurd Conclusion

7 Impersonal Average Principle (Maximize average utility)

Has no value

Is the only form of value

Refuted by the Two Hells (406) and by Hell Three (422)



Theories 1 and 2

Theory 1:  Impersonal Total Principle:  Maximize total utility.




The Repugnant Conclusion:  For any possible population of at least ten billion people, all with a very high quality of life, there must be some much larger imaginable population whose existence, if other things are equal, would be better, even though its members have lives that are barely worth living.


This conclusion must be rejected.   Therefore the notion of maximizing the totality of what makes life worth living is not theory X.


This same conclusion refutes:

Theory 2:  Quality has value, but a loss of quality can always be outweighed by gain in quantity

One solution to this problem is to maximize not total but average utility.  This, however, creates new problems:


Theory 7

Theory 7: The Impersonal Average Principle:  If other things are equal, the best outcome is the one in which people's lives go, on average, best.


This seems to imply a policy of population reduction, but it is refuted by:


The Two Hells (406):

Hell 1:  Last generation consists of 10 people who suffer terribly for 50 years

Hell 2:  Last generation consists of 10,000,000 people who suffer just as much for 50 years minus a day.

On the impersonal average principle, Hell 2 is to be preferred, but this is absurd


This can also be refuted by the Policy of Culling (which Parfit does not discuss)


Theories 5 and 6  (Theories of enough)


Theories 5 and 6 embody what Parfit considers to be a desirable asymmetry, setting limits positive but not negative quantity.  That is, once there is enough quantity there is no moral reason to increase it.


This is motivated by the idea that it is wrong knowingly to have a wretched child but not wrong knowingly not to have a healthy child.


Theory 5 (Diminishing Returns):  Each additional unit of quantity counts less, and there is an upper limit of quantity which the total approaches asymptotically but never reaches.


Theory 6: (Limited Quantity View):  There is an achievable upper limit of quantity beyond which it does not count at all.


Both views are refuted by:

The Absurd Conclusion:  Suppose the limit is ten billion.  Consider

1.     a huge population in which one person in 10 billion has a life of unmitigated misery.  The rest are happy.

2.     a collection of populations of 10 billion each on different planets in which one person in every 10 billion has a life of unmitigated misery while the rest are happy.

On asymmetrical views (views which limit positive but not negative quantity—i.e., views 5-6) the first outcome is good because quantity is maximized, the second bad because of the ever-increasing total of misery.


Conclusion:  Asymmety is false.  Since we cannot limit the badness of unmitigated suffering, there can be no limit to quantity's positive value.


This leads us to theories that do not limit quantity of value but still count quality:

Theories 3 and 4


Theory 3 (Appeal to the Valueless Level):  Quantity has no upper limit, but lives whose quality is below a certain level don’t count, even if such lives are worth living


Theory 4 (The Lexical View):  Quantity has no upper limit and every life worth living counts toward the positive total, but no amount of mediocre lives could equal the value of one blissful life.  (cf. Superman and herd)


Variants of Absurd and Repugnant Conclusions

Appeal to Valueless Level Implies variants of the absurd and repugnant conclusions:

(A)     If one in 10 billion had a life not worth living, and the others had lives worth living but just below the valueless level, this would be worse than if there were nobody at all. (Positive value doesn't count below the valueless level.)

(R)     If there were 10 billion people with a high quality of life, then there could be a much larger population whose existence would be better, even though its members all have lives barely above the valueless level.

If we replace the word 'valueless' by 'mediocre', the same conclusions follow from the Lexical View.


Hence none of the seven views seems to work.


Views 3 and 4 are further refuted by:




The problem that sets up the Mere Addition Paradox is the claim made by most averaging views that it is worse to add people whose quality of life is slightly below average, no matter how high average is. 


This Parfit denies.


In the diagram below, height represents quality and width quantity.


As we move from A+ to Divided B, the least well off improve more than the best well off lose.


In Theory 3 (Appeal to the Valueless Level) A is better than B:

Just assume lives in B are below the level at which quantity has value.


In Theory 4 (Lexical View) A is better than B: 

Since quality is about 4/5 as high in B as in A, it cannot be both that lives in A are above the blissful level and lives in B below the mediocre level.  So if B's population was large enough, B could be better than A.   But B's population is only twice A's.


Hence on either view A is better than B


The Mere Addition Paradox (First Version):

'<' means "is worse than"

          1        B º divided B                  Obvious

          2        A+  <  Divided B            A+ has lower average quality

                                                          [also considerations of fairness]

So     3        A+  <  B                         1,2

Let     4        B < A                             Appeal to Valueless Level, Lexical

                                                          View  (Quantity does not always


So     5        A+  <  A                         3,5

But Parfit thinks this is absurd.  Mere addition can't hurt


Since Parfit accepts 1 and 2, he rejects 4 and concludes ~(B < A).


It may seem as if this conclusion iterated implies the Repugnant Conclusion, but it doesn't, because it is not the case that:

                   (B is not worse than A) ® (A is at least as good as B)

i.e. it is not the case that:

                   ~(B < A) ® (B £ A)

Hence "not worse than" is not transitive and does not lead us to Repugnant Conclusion.


Mere Addition Paradox (Second Version):

Unlike the first version of the paradox, the second gets us something very much like the Repugnant Conclusion:



Each block represents 10 billion

Populations in New A…New Z are equal and of many 10s of billions.

Population of A+ is 20 billion.


Iterating the move from New A to New B eventually gets us to New Z where  lives are just above the bad level indicated by horizontal line. This is a form of the Repugnant Conclusion.


The assumptions here are:

          1        A+ < New A                   Improvement outweighs loss in

                                                          initial worlds; moreoever

                                                          Mere Addition can't hurt (434)

          2        New A < New B              Maximin, Equality, Beneficence

          3        New B < New C              Maximin, Equality, Beneficence

          4        New C < New D             Maximin, Equality, Beneficence




          N       New Y < New Z              Maximin, Equality, Beneficence

So     (N+1) A+ < New Z                             1-N


(N+1) is the:

New Repugnant Conclusion:  In A+ there are two groups of 10 billion people, one of which has a quality of life far higher than anyone who has ever actually lived and the other of which has a quality even higher.  In New Z there is a much larger number of people none of whom has a quality of life much above the bad level. New Z is better than A+.


The third version of the paradox is the same idea with each block representing 1000 years.


It seems much less repugnant that way.