Medical Hypotheses (2007) 68, 184–187

http://intl.elsevierhealth.com/journals/mehy

Non-reciprocal altruism may be attributable to

hyperbolicity in social discounting function
Taiki Takahashi *

Department of Cognitive and Behavioral Science, The University of Tokyo, 3-8-1 Komaba, Meguro,
Tokyo 153-8902, Japan

Received 14 May 2006; accepted 16 May 2006

Summary Humans show altruism even for chance acquaintances whom they will never meet again (non-reciprocal
altruism). With respect to evolutionary and economic perspectives, reciprocal altruism is not actually puzzling,
because reciprocal altruism maximizes subjects’ benefit/fitness in the long run; while non-reciprocal (pure) altruism
still presents a challenge to evolutionary biology and neuroeconomics. Understanding neuropsychological bases of non-
reciprocal altruism/generosity for unrelated people is also important for neuropsychiatry, because several types of
impulsive psychiatrics (e.g., social phobia, substance abusers, psychopaths) often have reduced altruism/generosity
for other people, which is associated with their problematic/impulsive social behavior. Regarding the relationship
between impulsivity and reciprocal altruism, it has been shown that subjects who are impulsive in intertemporal choice
(having a large time-discounting rate) have low degrees of reciprocal altruism, which is in line with economic theories
and neuropharmacologically treatable with serotonin/norepinephrine reuptake inhibitors. However, little is known
regarding the relationship between non-reciprocal altruism and discounting behavior, and neuropsychological
processing underlying the relationship. Recently, Jones and Rachlin reported that generosity for other individuals with
different social distances follows a hyperbolic discounting function, which has been employed to describe inconsistent
time-discounting. In this study, I propose that non-reciprocal (pure) altruism may be, at least partially, attributable to
hyperbolicity of a social discounting function (inconsistency in ‘‘interpersonal choice’’). Neuropsychological
mechanisms underlying interpersonally-inconsistent social discounting and possible implications for neuropsychophar-
macological treatments for impulsive psychiatrics’ problematic social behavior are discussed.
c 2006 Elsevier Ltd. All rights reserved.

‘‘A chance acquaintance is a divine ordinance.’’ Background

(Proverb).
Altruism

Axelrod and Hamilton’s classical work on evolution

* Tel./fax: +81 3 5454 6709.
E-mail addresses: taiki@ecs.c.u-tokyo.ac.jp,
hotmail.com.
of social cooperation has demonstrated that coop-
erative behavior (Tit-for-Tat) evolves under the
Iterated Prisoner’s Dilemma Game condition (evo-
taiki103@ lution of reciprocal altruism), because reciprocal
altruism maximize subject’s total benefit/fitness


0306-9877/$ - see front matter c 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mehy.2006.05.057
Non-reciprocal altruism may be attributable to hyperbolicity in social discounting function
integrated/summed over time [1]. Hence, reciprocal
altruism is not purely altruistic, but self-controlled
(i.e., far-sighted) selfish behavior. Psycho-pharma-
cological studies reported that impulsive psychiat-
rics’ anti-social behavior is associated with their
reduced self-control in intertemporal choice [2].
The relation between intertemporal choice and
altruism has been studied in psychology, behavioral
econology, and neuroeconomics [3]. For instance,
neuropsychopharmacological studies reported that
subjects with impulsivity in intertemporal choice
(i.e., large temporal discounting rate, discussed
later) are less cooperative/altruistic [4,5] and
inactivation of brain serotonergic activity reduces
cooperation/reciprocal altruism in the Iterated
Prisoner’s Dilemma Game [6].
In contrast to reciprocal altruism, the existence
of pure altruism, i.e., non-reciprocal altruism, is
still puzzling and yet to be more extensively stud-
ied, because non-reciprocal altruism does not in-
crease subject’s fitness/benefit even in the long
run. This means that, in non-iterated, single-shot
social interactions, even a patient (in terms of
temporal discounting) subject should, normatively
speaking, defect the other subjects when they
are mere chance acquaintance, because the sub-
ject cannot expect reciprocal rewarding from the
mere chance acquaintances in the future. Several
neuroeconomic and evolutionary biological studies
have investigated why people are altruistic and
generous even for unrelated individuals [7], but
the final answer is still unknown.
Before answering this difficult question, it may
be helpful to examine the neural or psychophysical
bases of altruism/social generosity in a mathemat-
ically rigorous manner. Regarding the mathemati-
cal characteristics of the relation between
altruism and social distance, Jones and Rachlin re-
cently found that generosity for other people with
different social distances are well described by a
hyperbolic function, which has been frequently
employed for the description of subjects’ inter-
temporal choice [8]. Therefore, let us start from
reviewing mathematical characteristics of tempo-
ral discounting functions briefly.
Intertemporal choice
Discounting of delayed rewards refers to the obser-
vation that the value of a delayed reward is dis-
counted (reduced in value or considered to be
worth less) compared to the value of an immediate
reward [9]. Studies in psychopharmacology, psychi-
atry, behavioral neuroscience, and economics have
been focused on how subjects discount delayed
185
rewards. Notably, it has repeatedly been demon-
strated that substance abusers/drug addicts more
steeply discount delayed rewards than non-drug
dependent subjects [10].
Samuelson’s classical economic theory, has as-
sumed that individuals discount delayed reward in
a consistent manner [9]. This type of discounting
(exponential discounting) follows the exponential
equation
VðDÞ ¼ Vð0Þ expðkDÞ; ð1Þ
where V is the subjective value of a reward, V(0) is
the (objective) amount of the reward, and D is the
length of delay until the delivery of reward. The
free parameter k, independent of D, is an index
of the degree of discounting, i.e., larger k values
correspond to steeper delay discounting. The con-
sistency in intertemporal choice requires a dis-
counting rate (DR) defined as Œ(dV/dD)/V Œ (=k in
Eq. (1)) is constant over time (independent of D)
[9,10]. However, subsequent empirical studies
have repeatedly observed anomalies and inconsis-
tencies in subjects’ intertemporal choice behavior
[9,10]. For instance, subjects are more impulsive
in the near future (i.e., large discounting rate) than
in the far future (i.e., small discounting rate),
which has been described with the hyperbolic dis-
counting function
VðDÞ ¼ 1=1 þ kD; ð2Þ
resulting in inconsistency in intertemporal choice
(i.e., Œ(dV/dD)/V Œ is a decreasing function of D,
in contrast to Eq. (1), in which Œ(dV/dD)/V Œ = k is
independent of D) [9,10]. This inconsistency results
in preference reversal over time and loss of self-
control in intertemporal choice often observed in
impulsive psychiatrics such as drug addicts [10].
Hypothesis
Let us again remember that reciprocal altruism is a
type of selfish behavior, in that reciprocal altruistic
behavior maximizes benefit/fitness of subjects in
the long run. Hence, normatively speaking, subjects
with small time-discounting rates (patience in inter-
temporal choice) should behave in a reciprocal
altruistic manner [3]. Actually, several studies
reported that subjects’ time-discounting rates are
associated with their cooperative tendency under
Iterated Prisoner’s Dilemma Game condition [3–
5].This again suggests that reciprocal altruism is a
type of self-controlled selfish (self-interest-maxi-
mizing) behavior, which is not contrary to evolution-
ary theory and economics’ assumption of
selfishness. The selfish, reciprocal altruism moti-
vates subjects to cooperate with other subjects
186
who are genetically related or socially close to the
subjects [1,3,7]. However, as stated, non-reciprocal
altruism, often observed in humans, is pure altruism
in the sense that it will not pay even in the long run
[7]. The existence of non-reciprocal altruism in hu-
mans is therefore puzzling. Recently, Jones and
Rachlin [8] demonstrated that declining of generos-
ity/altruism, according to an increase in the social
distance of the other subject, follows the hyperbolic
discounting function with social distance N
AðNÞ ¼ Að0Þ=1 þ sN; ð3Þ
where N is an ordinal social distance of the other
subject (for the self, N = 0), A(N) is altruism/gener-
osity towards the subjects with social distance N,
and s (>0) is a free parameter independent of N.
The parameter s is a selfishness parameter, because
larger s values corresponds to steep decrease in gen-
erosity as social distance increases. Note that the
potential benefit from possible social interaction
between the self and the other subject will de-
crease, as the social distance between the self and
the other increases, as the economist Akerlof states
[11]. It is noteworthy that Jones and Rachlin have
further observed that subject’s social discounting
behavior is better described by the hyperbolic equa-
tion, in comparison to an exponential function
AðNÞ ¼ Að0Þ expðsNÞ; ð4Þ
indicating that there may be inconsistency in inter-
personal allocation of resources. Mathematically,
social discounting rate SDR(N) can be defined as
SDRðNÞ :¼ jðdA=dNÞ=Aj: ð5Þ
It can be said that this social discounting rate is a
parameter indicating the selfishness (non-altruism)
at social distance N.
In exponential social discounting, SDR(N) = s and
(d/dN)SDR = 0 (consistent social discounting);
while in hyperbolic social discounting,
SDRðNÞ ¼ s=ð1 þ sNÞ
ðhyperbolic social discounting rateÞ: ð6Þ
Because
2
ðd=dNÞSDRðNÞ ¼ s=ð1 þ sxÞ < 0
ðnote that s > 0Þ; ð7Þ
the hyperbolic social discounting rate is a decreas-
ing function of social distance N (inconsistency in
interpersonal choice). This indicates that steepness
of social discounting (degree of selfishness for
N = 0, a tendency of selfish nepotism for N > 0) is
smaller at large social distance than at small social
distance. Namely, subject’s selfishness parameter
decreases as social distance increases, which will
never occur in exponential discounting. This results
Takahashi
in enhanced altruism to unrelated subjects (non-
reciprocal altruism) in hyperbolic social discount-
ing, in comparison to consistent interpersonal
choice (exponential social discounting). In other
words, as mathematically be derived, subjects’
selfishness/greed towards others with large social
distance (unrelated subjects) may be smaller
(more altruistic) than expected from a rational so-
cial choice (i.e., exponential social discounting). It
should be stressed here that in Jones and Rachlin’s
demonstration, social distance is not a cardinal but
ordinal number; namely, social distance of a com-
pletely unrelated other subject may not be infinite
(large but still finite), therefore, altruism for
chance acquaintances following the hyperbolic dis-
counting function may still be non-zero (>0) and
stronger than expected. Considering that subjects
will never meet the mere chance acquaintance,
this hyperbolic effect may result in non-reciprocal
altruism; i.e., marked altruism which cannot be ex-
plained or justified in terms of self-interest be-
cause the other’s social distance is too large to
expect future social interactions and associated
reciprocation from him/her. Furthermore, it
should be noted that if subjects are not capable
of perceiving social distances of others in a ordinal
manner, it is obviously impossible for them to so-
cially discount resources hyperbolically (or expo-
nentially). Therefore, paradoxical altruistic
behavior might be observed only in animals that
are capable of perceiving ordinal social distance.
This may explain why altruistic behavior is not fre-
quently observed in the animal kingdom [3]. With
respect to impulsive social behavior in psychiatrics,
even when psychiatrics show impulsive social
behavior towards their close acquaintances (e.g.,
domestic violence), it is possible, from the present
hypothesis, that they are less impulsive/greedy to-
wards unrelated people than expected.
Implications for treatment for impulsive
social behavior in psychiatrics
It is still unknown whether impulsive psychiatrics
(e.g., drug addicts and psychopaths) are less altru-
istic towards completely unrelated others than
healthy controls, although it has been demon-
strated that subjects with strong impulsivity in
intertemporal choice show less reciprocal altruism
[4,5]. Reciprocal altruism can possibly be en-
hanced by neuropharmacological treatments which
enhance patience/self-control in intertemporal
choice (e.g., serotonin/norepinephrine reuptake
inhibitors) [6,12,13]; on the contrary, little is
known regarding how social discounting rate can
Non-reciprocal altruism may be attributable to hyperbolicity in social discounting function
be modified by neuropsychopharmacological treat-
ments. Previously, I proposed that inconsistency in
intertemporal choice may be due to time-percep-
tion error following Weber–Fechner law [14].
Whether the inconsistency (hyperbolicity) in social
choice is also attributable to an error in the
perception of social distance is yet to be eluci-
dated, because little is known about neurobiologi-
cal correlates of social distance perception. One
promising candidate is amygdala-HPA (hypotha-
lamic pituitary axis) activation, induced by social
interaction, because the presence of social sup-
port by subjects with small social distance (close
friends) dramatically reduces social stress-induced
amygdala-HPA activation [15]. Several studies fur-
ther indicated that oxytocin administration (known
to reduce social stress-induced cortisol elevation
[15]) increases non-reciprocal altruism [16], which
is consistent with our previous observation that
social stress-induced cortisol elevation was nega-
tively associated with subjects’ tendency of trust-
ing other people in general [17,18]. Therefore,
future studies should examine how neuropsycho-
pharmacological manipulation changes subjects’
perception of social distance and social discount-
ing rate, in order to establish better treatments
for social phobias and impulsive psychiatrics such
as substance misusers and psychopaths.
Acknowledgements
The research reported in this paper was supported by
a grant from the Grant-in-Aid for Scientific Research
(‘‘21st century center of excellence’’ grant and
Grant #17650074) from the Ministry of Education,
Culture, Sports, Science and Technology of Japan
and Yamaguchi Endocrinological Disorder Grant.
References
[1] Axelrod R, Hamilton WD. The evolution of cooperation.
Science 1981;211(4489):1390–6.
187
[2] Petry NM. Discounting of delayed rewards in sub-
stance abusers: relationship to antisocial personality
disorder. Psychopharmacology (Berl) 2002;162(4):
425–32.
[3] Stevens JR, Hauser MD. Why be nice? Psychological
constraints on the evolution of cooperation. Trends Cogn
Sci 2004;8(2):60–5.
[4] Yi R, Johnson MW, Bickel WK. Relationship between
cooperation in an iterated prisoner’s dilemma game and
the discounting of hypothetical outcomes. Learn Behav
2005;33(3):324–36.
[5] Harris AC, Madden GJ. Delay discounting and performance
on the prisoner’s dilemma game. Psychol Rec 2002;52:
429–40.
[6] Wood RM, Rilling JK, Sanfey AG, Bhagwagar Z, Rogers RD.
Effects of tryptophan depletion on the performance of an
iterated Prisoner’s Dilemma game in healthy adults. Neu-
ropsychopharmacology 2006;31(5):1075–84.
[7] Fehr E, Rockenbach B. Human altruism: economic, neural,
and evolutionary perspectives. Curr Opin Neurobiol
2004;14(6):784–90.
[8] Jones B, Rachlin H. Social discounting. Psychol Sci
2006;17(4):283–6.
[9] Frederick S, Loewenstein G, O’Donoghue T. Time discount-
ing and time preference: a critical review. J Econ Lit
2002;40:351–401.
[10] Bickel WK, Marsch LA. Toward a behavioral economic
understanding of drug dependence: delay discounting
processes. Addiction 2001;96:73–86.
[11] Akerlof GA. Social distance and social decisions. Econome-
trica 1997;65(5):1005–27.
[12] Tse WS, Bond AJ. Serotonergic intervention affects both
social dominance and affiliative behaviour. Psychopharma-
cology (Berl) 2002;161(3):324–30.
[13] Tse WS, Bond AJ. Reboxetine promotes social bonding in
healthy volunteers. J Psychopharmacol 2003;17(2):
189–95.
[14] Takahashi T. Loss of self-control in intertemporal choice
may be attributable to logarithmic time-perception. Med
Hypotheses 2005;65(4):691–3.
[15] Heinrichs M, Baumgartner T, Kirschbaum C, Ehlert U. Social
support and oxytocin interact to suppress cortisol and
subjective responses to psychosocial stress. Biol Psychiatry
2003;54(12):1389–98.
[16] Kosfeld M, Heinrichs M, Zak PJ, Fischbacher U, Fehr E.
Oxytocin increases trust in humans. Nature 2005;435(7042):
673–6.
[17] Takahashi T, Ikeda K, Ishikawa M, Kitamura N, Tsukasaki
T, Nakama D, et al. Interpersonal trust and social stress-
induced cortisol elevation. Neuroreport 2005;16(2):
197–9.
[18] Takahashi T. Social memory, social stress, and economic
behaviors. Brain Res Bull 2005;67(5):398–402.