Amagire de Sine

8/12/2019 Amagire de Sine

1/30

Finite Rational Self-Deceivers

D. S. Neil Van Leeuwen

Stanford University, Department of Pilosopy

Abstract: I raise three puzzles concerning self-deception: (i) a conceptualparadox, (ii) a dilemma about how to understand human cognitive

evolution, and (iii) a tension between the fact of self-deception and

Davidsons interpretive view! I advance solutions to the first two and la" agroundwor# for addressing the third! $he capacit" for self-deception, I

argue, is aspandrel, in %oulds and &ewontins sense, of other mental

traits, i!e!, a structural b"product! $he iron" is that the mental traits of

which self-deception is a spandrel'b"product are themselves rational!

!ntroduction" a Parado#, a Dilemma, and a $ension

$hree puzzles plague the notion of self-deception!

irst, the classicparadoxof self-deception points to an apparent incoherence in

the ver" concept! If A deceives , then A does not believe what she ma#es believe! *o

if A deceives A, A must not believe what A believes! $hus the existence of self-deception

seems to entail a contradiction! ut manifold examples of human behavior are

appropriatel" called cases of self-deception! *o how is self-deception possible+

*econd, #nowledge isprima faciecritical to the evolutionar" success humans

have had! ur vision, hearing, other senses, and higher cognitive abilities largel" seem to

owe their existence to the fitness value of #nowledge! $he widespread human capacit"

for self-deception, however, undermines #nowledge! $he prominence of self-deception,

therefore, creates a dilemmaabout how to understand the fitness value of #nowledge in

general! If #nowledge enhances fitness, then the capacit" for self-deception shouldnt

exist since it undermines #nowledge! ut it does exist! If #nowledge does notenhance

fitness, then the existence of self-deception is not a problem for evolutionar" theor"! ut


2/30

we are then deprived of the most obvious explanation of our most complicated and

interesting traits! .ow shall we solve this dilemma+

$he third puzzle is motivated b" the wor# of Donald Davidson, who indeed

recognized it: the widel" held interpretive view of the mental holds that rationalit" is

constitutive of the mental, including beliefs!/ne cannot sensibl" attribute beliefs to an

agent unless those beliefs ma#e rational sense to the person attributing them! ut self-

deception is an irrational belief state that people, with apparent 0ustification, attribute

often! $hus there is a tensionbetween holding that rationalit" is constitutive of belief and

holding that belief can come about b" self-deception! .ow shall we resolve this tension

between the force of the interpretive view and the fact of self-deception+1

In this paper, I advance solutions to the first two puzzles! I also, b" explaining

how self-deception relates to rational capacities and evidence, la" a groundwor# for

addressing the third! I do this b" advocating three theses!

! I offer a new definition of self-deception 2$hesis 3, one that captures the

epistemic tension inherent in sa"ing one deceives oneself, while avoiding the

absurdit" of sa"ing one believes what one doesnt! I state the definition full" insection !

/! I advance the byproduct viewof self-deception 2$hesis /3, which runs as follows:The capacity for self-deception is a byproduct of a number of critical abilities

humans have that enable us to cognize and behave rationally given finite minds!

*elf-deception, in short, is a b"product of finite rationalit"!

$he b"product view ta#es a s"nchronic perspective on the capacit" for self-deception, but

it also ma#es salient a diachronic explanation of that capacit"!

1! I advance the view 2$hesis 13 that the capacit" for self-deception is what

biologists *teven 4! %ould and 5ichard &ewontin call aspandrel! $he thesis,precisel", is: The capacity for self-deception is a structural byproduct of features

that were not selected for their role in the production of that capacity! I do not

claim here that the features I cite are themselves adaptations6 it is enough for m"

Finite Rational Self-Deceivers - 2


3/30

view that there is independent 0ustification for positing their existence apart from

their role in self-deception!

$heses / and 1 are identical as to the features the" claim the capacit" for self-deception is

a b"product of, but the" differ as to the claims the" ma#e aboutthose features! $hesis /

claims that the" are critical to finite rational cognition and behavior! $hesis 1 claims that

the" did not arise in natural selectionforproducing self-deception!

$he definition is meant to solve the paradox! $heses / and 1, the b"product and

spandrel theses, solve the second puzzle in a wa" that does not force one to den" the

fitness value of #nowledge! As a spandrel, the historical existence of the capacit" for self-

deception rode pigg"-bac# on the evolution of those features of which it is a spandrel!

*ince those features are rational, we can explain the capacit" for self-deception as the

downside of a tradeoff: selection favored a pac#age of features that are themselves

rational, but have an irrational b"product! $he whole pac#age will then have to have been

more fit than competing pac#ages, but it will be the rationalit" of it that contributes to the

fitness7not the self-deception!

It should be apparent that the b"product view is of independent philosophical

interest! 5ationalit" is of course a contested concept, but for clarit" I offer the following

wor#ing definition: a capacit" is rational if it is (a) conducive to truth and coherence in an

agents belief set or (b) conducive to means-end coherence in practical planning, and (c)

is not directl" detrimental to either (a) or (b)8! $he cluster of features I identif" constitute

one wa" of having rationalit" in a finite mind with practical pro0ects, but that cluster

"ields the capacit" for self-deception as a b"product! Is there some other wa" of having

rationalit" in a finite mind+ 9" discussion will suggest that, for an" finite creature with

human-li#e desires, rationalit" brings with it the capacit" for self-deception!

Finite Rational Self-Deceivers -


4/30

I structure this paper as follows! In section , I present m" definition of self-

deception and explain how it resolves the paradox! In section /, I defend m" theses about

the capacityfor self-deception, arguing for both theses in parallel! I then describe, in

section 1, a paradigm case of self-deception and appl" m" framewor# to it! $o explore

the wider implications m" theor" has, I argue in section 8 that the spandrel view of self-

deception challenges the dominant adaptationist paradigm in evolutionar" ps"cholog"!

*ection concludes with implications about self-deception and human rationalit"!

% Definin& Self-Deception and Resolvin& te Parado#

n the wa" to defining self-deception, I start with three paradigm cases!

Imagine a college dropout whos aware of several important pieces of evidence

that suggest finishing his degree will improve his 0ob prospects! .es seen statistics on

earnings6 his brother who graduated has fared much better at getting 0obs6 and several

positions he wants re;uire a degree! ut he wants it to be the case that his chances are

good even without finishing6 he becomes self-deceived that the" are!


5/30

for most of m" discussion as a

general wa" of referring to this motivational component!$hird, the doxastic alternative

is the proposition that the agent in some senseshouldbelieve! $he doxastic alternative for

the college dropout is:finishing my degree is needed to improve my #ob prospects!

If we return to the paradox, we find two ma0or classes of views aimed at defining

self-deception in a wa" that resolves it! irst, theres the class of views that holds the

product of self-deception is a belief6 when Im self-deceived thatp, I believe thatp! $he

paradox is then resolved either b" drawing a division in the mind between the deceived

part (which believesp) and deceiving part (which believes $p), or b" den"ing the

re;uirement that the self-deceiver also believes $p!?$he second class of views holds that

the product of self-deception is an avowal or avowed belief, a mental state underl"ing

verbal behavior, but lac#ing important properties of genuine beliefs, such as deep

connections to actions aside from merel" verbal ones! $he belief view is more common,

and is held b", among others, Davidson (@@), &azar (@@@), 9ele (/BB), Cears (@8),

$albott (@@), and m"self! ariants of the avowal view are held b" Audi (@), 5e"

(@), and un#houser (/BB)!

Finite Rational Self-Deceivers - %


6/30

$he avowal view seems to provide an easier solution to the paradox: the self-

deceiver doesnt both believethatpand believe that $p, since the wa" in whichpis held

is merel" an avowal!

$he problem with the avowal view is that it resolves the paradox b" den"ing deep

connections between self-deception and action! ut consider a gambler who marches to

the casino self-deceived that he can climb his wa" out of debt6 hes actingon his self-

deception! I hold the belief view, because the self-deceived gambler wants to get out of

debt and believesthat he can do so b" gambling! &i#ewise, thellos self-deception

explains not merel" his verbal behavior, but also his actionof #illing his wife! $he

product of self-deception pla"s the same role in action explanation as other beliefs, so it

is best viewed as a belief!E

&et us then hold fixed that the product of self-deception is a belief, and see what

our paradigm cases have in common!

irst, the agents arent forming the self-deceptive belief out of lac# of intelligence

or relevant evidence6 the belief goes contraryto their epistemic norms (rational rules for

belief formation) and the evidence the" have! If the soccer pla"er had similar evidence

about someone elses coach, she would correctl" conclude that coach was the angr" t"pe!

*econd, a desire, with content related to the belief formed, is ma#ing the causal

difference in the belief formation process! $hat the desire ma#es the causal difference

will be clear from a thought experiment6 suppose thello 0ust didnt care whether

Desdemona was faithful! $hen, I thin#, his self-deceptive belief wouldnt have arisen! n

the basis of these considerations, I give this definition 2$hesis 3:

Finite Rational Self-Deceivers - &


7/30

'n agent is in a state of self-deception if and only if

(i) she holds a belief*

(ii) that belief is contrary to what her epistemic norms in con#unctionwith what evidence she has would usually dictate* and

(iii) a desire* with content appropriately related to the belief formed*

causally ma+es the difference to what belief is held in anepistemically illegitimate fashion!@

=Fpistemicall" illegitimate> here is to be understood as illegitimate relative to the

epistemic norms that the agent actually has, since believing contrar" to norms one

doesnt have shouldnt count as self-deception (although perhaps it would be wishful

thin#ing)!BA self-deceived agent, I hold, does not 0ust violate standards of rationalit"6

she violates her own standardsof rationalit"!

9" definition is meant to resolve the

paradox of self-deception as follows: it captures the epistemic tension/involved in self-

deception b" re;uiring that the agent have compelling evidence to the contrar" of the

belief formed, but b" not re;uiring both beliefs be held it escapes positing the

ps"chological absurdit" (if it is absurd) of holding two contradictor" beliefs!

Ghat relation must obtain between the content of the deceptive element (the

desire) and the product of self-deception (the belief)+ $here are three possibilities and

thus three main t"pes of self-deception! irst, the content of the desire can be thesameas

the content of the resulting belief6 I call this wisful self-deception, or self-deception

thats continuous with wishful thin#ing! $he college dropout, for example, has wishful

self-deception! .e believes what he wants to be the case! $his #ind of self-deception isnt

planned6 one slides into it! *econd, the desire can be to have the beliefthats ultimatel"

formed, as with the nervous soccer pla"er6 I call this willful self-deception! Cascal, it

seems, was willful! inall", the content of the desire can be contrar" to the content of the

resulting belief, as in thello6 I call this dreadful self-deception1!8

Finite Rational Self-Deceivers - ,


8/30

$he chart below summarizes forms of self-deception and related phenomena! Dip

means i desires thatp6 'ip means i has evidence that b" her epistemic norms on the whole

favors believingp! $hese are all different forms of irrationalit" agent i can have in

believing thatp ((ip)! $he far left column refers to the deceptive element!

Non-self-deceptive )elief Self-deceptive )elief

*'ip, *'i*p, (ip 'i*p, (ip

Dip wishful thin#ing wishful self-deception

Di(ip willful thin#ing willful self-deception

Di*p dreadful thin#ing dreadful self-deception

$he idea behind the wishful thin#ing, willful thin#ing, and dreadful thin#ing column is

that the agent, with desire pla"ing a causal role, forms a belief without having compelling

evidence that its true or compelling evidence its false! $he various forms of self-

deception are stronger forms of irrationalit", for in those cases the agent has evidence that

supports the belief that runs contraryto the belief thats formed7in fact, evidence that

supports the negation of whats believed! Ge can see, then, that there is a continuum

between wishful thin#ing and self-deception, with the line being crossed when the

evidence contrar" to the product is sufficient b" the agents epistemic norms to support

the doxastic alternative!

In what follows, I assume m" definition and conceptual framewor# are plausible

and address what the capacit" for self-deception consists in and wh" it exists, focusing on

wishful self-deception, which seems the most common form!

+ $e 'tiolo&y of Self-Deception

$he second puzzle of self-deception, the dilemma about the fitness value of

#nowledge, raises the ;uestion of wh" the capacit" for self-deception exists!

Finite Rational Self-Deceivers -


9/30

.rima facie, the capacit" for self-deception loo#s bad for fitness, so its puzzling

that it hasnt been selected out! 9an" examples of self-destructive behavior caused b"

self-deception support the intuition that it has negative fitness value: people get ta#en

advantage of in relationships the" self-deceptivel" believe are good, get in fights the"

cant win but self-deceptivel" believe the" can, fail to fix serious problems the"re self-

deceived about, and so on!


10/30


11/30

arises! I show that the operation of the features in this complex to produce self-deception

is wholl" intelligible without appeal to an" mechanism whose =purpose> (evolutionaril"

or otherwise) is to produce self-deception! I then highlight other features that aid the

process of self-deception6 these can be divided into epistemicfacilitators, factors that in

certain contexts ma#e it easier to believe contrar" to usual epistemic norms, and desire

sources, aspects of mind that give rise to the sorts of desires that easil" figure in self-

deception! $hroughout I will ma#e the case that the features identified are conducive to

rational cognition, goal-setting, and goal attainment given finite limits! If this is true, then

the capacit" for self-deception is not merel" a ;uir# that arises out of unimportant

features of what our brains are li#e6 its rather a capacit" that arises from features

essential to human rationalit"!E

+.% $e 'ssential omple#

*even features of the human mind comprise the essential complex that "ields self-

deception! $he first three are:

'%:Desires have a characteristic sting that accompanies anticipation or evidence oftheir non-satisfaction

'+:3umans have the ability to attend selectively to inputs and evidence

':3umans have a general inclination to avoid discomfort@

.ow do these features of mind contribute to self-deception+ *a" I desire thatp!

$hus I also have the disposition to feel a sting when $pis anticipated! ecause of this

desire thatp, when I encounter evidence of $p, I have a feeling of discomfort or

;ueasiness (FH)! $his evidence must be cognized as evidenceof $pin order for it to

cause the discomfort7otherwise it simpl" wouldnt bother me! $his discomfort is there

especiall" when m" attention is on the $pevidence, but abates when m" attention shifts



12/30

7especiall" when it shifts (FH/) topevidence (what little there ma" be)!/BAgainst a

bac#ground inclination to avoid discomfort (FH1), m" attention shifts to thepevidence!

*elf-deception ensues, for the focus of attention on thep evidence gives rise to thep

belief, even though the totalevidence possessed in con0unction with epistemic norms

would dictate believing $p!

$hus runs the basic causal chain from the desire that p to the belief thatp, despite

compelling evidence that $p! Importantl", it isnt the case that the normal functions/of

FH-FH1 are to produce self-deception! FH, the sting of desires, is conducive to goal-

attainment6 it prods us out of situations in which our desires are unli#el" to be satisfied!

FH/, selective attention, is an essential feature of an" finite cognitive s"stem with

interests6 without it the mind would be mired in a wash of inputs, most of which would be

irrelevant to the organisms ends! FH1, general inclination to avoid discomfort, t"picall"

has the function of #eeping one out of situations that ma" be harmful and is thus

generall" conducive to goal attainment! $he iron" of FH1 is that, b" helping ma#e self-

deception possible, it sometimes contributes to #eeping one inharmful situations about

which one is self-deceived!//

$he essential complex also includes:

':4vidence in the human mind is structurally organized

Ge wouldnt be able to attend to evidence selectivel" if that evidence werent organized

in wa"s that allow us to search through it! $he structure of the Fvidence ox, so to spea#,

is essential to our abilit" to cognize rationall", for without it selective attention would be

aimless! ut it also is an essential component in our capacit" for self-deception! $hus, to

relate this point explicitl" to $hesis /, we have the surprising result that self-deception is,



13/30

at least in part, a b"product of rationalcapacities! (" wa" of contrast, an infinite rational

being wouldnt need evidence in her mind to be structurall" organized, for infinite

powers would allow attending to everythingat once, with no need to search through the

evidence!)

$he essential complex, as Ive described it so far, presupposes the following

feature:

'/:3umans form beliefs on the basis of evidence in con#unction with epistemic norms

$he deep iron" is that this rational feature of mind is both subverted and implicated in the

same self-deception! Its subverted with respect to total evidence, but implicated in the

formation of the self-deceptive belief b" the evidence selectivel" attended to! /1

or completeness, I add:

'0"3umans experience pleasure at evidence that their desires will be satisfied

'1"3umans see+ pleasure

An example of FH? in action is the exuberance felt at the end of a sports competition

when victor" is in sight! It should be clear immediatel" how FH? and FHE contribute to

attainment of the goals that one has, but the" are also implicated in the comfort-driven

modulation of attention that produces self-deception!/8

+.+ Facilitators and Desire Sources

Afacilitatoris a mental feature that ma#es easier the failure of epistemic norms

and evidence in self-deception! A desire sourcegives rise to desires that are apt to cause

self-deception!


14/30

facilitators and (second) the most important desire sources in the capacit" for self-

deception!

F%" The web of belief has inertia

eliefs t"picall" occur in relation to a web/of other beliefs that 0ustif" them and

give them content! ne fact of human cognition is that the webs that constitute our belief

sets have inertia! A s"stem of beliefs does not easil" change entirel" due to the existence

of facts that are anomalous from the perspective of particular beliefs! $his aspect of our

cognitive econom" is largel" advantageous to coherence in our belief sets6 for if our webs

of beliefs underwent revolution with each discover" of anomalous fact, we would be in a

perpetual state of cognitive flux! It would not ma#e sense to perform a massive modus

tollenson our entire web simpl" because one belief encounters anomal"! $his explains to

some extent wh" confirmation bias, when it involves the tendenc" to loo# for

confirmation of things alread" believed, ma" be helpful6 it helps shield us from constant

cognitive revolution!

$he dar# side of the inertia of the web is that it facilitates self-deception! irst, it

ma#es it easier to hang on to beliefs under the influence of a desire even when the

evidence available has become compelling in the other direction! $he essential complex

ma" not alwa"s be sufficient b" itself to cause a cuc#old to hang on to the belief that his

wife is faithful against the weight of the evidence! ut given inertia of his alread"

existing web of beliefs, self-deception becomes possible! Juhn (@?/) discusses

scientists who wont adopt a new paradigm even when the rest of the scientific

communit" has signed on6 some of these are probabl" cases of inertia-aided self-

deception! urthermore, the inertia of the web helps explain self-deception in the teeth of

Finite Rational Self-Deceivers - 0"


15/30

strong evidence to the contrar" of the self-deceptive belief: multiple self-deceptions in a

context can underlie the conscious reclassification of evidence and holding of beliefs6 in

such cases the inertia of the web is the mental difference between a house of cards and a

house of cards with glue!

F+"Desires and other emotions generate con#ectures and thoughts

Antonio Damasio, inDescartes! 4rror(@@8), argues that two of the brains

emotional centers, the ventro-medial pre-frontal cortex and the am"gdala, are critical for

reasoning! .e doesnt differentiate between theoretical reason, reasoning to form true

beliefs, and practical reasoning, reasoning to form good actions6 his thesis is more

plausible when ta#en to be about practical reasoning! ut one of his points about the role

of emotions in life planning suggests how emotions can influence theoretical reasoning as

well! Fmotions, including desires, cause ideas to be suggested in our mind for

consideration! Ghen one desires approval, for example, thoughts of oneself being

approved of and accepted constantl" pass through consciousness! Fmotions themselves

cause the appearance of representations in thought (=images,> as Damasio puts it) of the

ob0ects at which the emotions are directed! $his, of course, facilitates planning and

practical reasoning about life choices, and Damasio shows that patients with damaged

emotional centers are incapacitated to act rationall": due to lac# of appropriate emotions

the right ideas often dont occur to them for consideration!5ut insofar as emotions and

desires propose thoughts* con#ectures* and hypotheses for the agent to consider* they will

also influence the formation of beliefs6theoretical reasoning! $his is simpl" because we

cant form beliefs about contents that havent occurred to us! $his mode of influence of

desires and emotions on belief formation is beneficial for both cognitive and practical

Finite Rational Self-Deceivers - 0%


16/30

goal-attainment, for without the guidance of emotion directing what we attend to our

cognitive capacities would be mired in consideration of an infinitude of useless

information!

Ge often have desires for things intangible but felt to be important, such as what

will happen in the future, the abilities of our children, the attitudes of friends and

enemies, and events we wont ever see! In short, we have desires for things about which

the beliefs fall far from the sensor" peripher" of the web of beliefs! Ghen such desires

thatparise, thep-thoughts arise also, as explained! $he perpetual recurrence, however, of

p-thoughts due to desires thatpcan be the first step in the self-deceptive formation of the

belief thatp6 for that same desire can, in virtue of the essential complex, cause selective

attention to evidence in favor of the belief thatp! *ince thep-thought is the first step in

the formation of thep-belief,the propert" the desire has of proposing representations for

thought, and of occup"ing the mind with them, thus ma#es it easier for self-deception to

occur!

F"3umans can apply differing degrees of s+epticism to different propositions

*#epticism7the withholding of belief from a proposition that does not meet a

certain level of 0ustification7comes in different degrees, ranging from refusing to

believe a newspaper account without hearing it corroborated elsewhere to deep Hartesian

s#epticism about the external world! 9ost humans can modulate the degree of s#epticism

applied in given circumstances6 citizens of a democrac", for example, will t"picall" appl"

a higher degree of s#epticism to information from a part" the" oppose than to information

from the part" the" support! eing s#eptical is a rational capacit", for it helps screen out

belief in falsehoods! ut appl"ing s#epticism differentiall" to different propositions can

Finite Rational Self-Deceivers - 0&


17/30

ma#e us less truth-trac#ing, not more, whenever the differential application of s#epticism

is caused b" desires other than the desire for truth! Discomfort with evidence in favor of

believing that $p, caused b" a desire thatp, can ma#e a person want to reclassif" that

evidence6 differential application of s#epticism can aid in reclassif"ing evidence and thus

in the self-deception thatp!

F"3umans can suppress unwanted memories

$he debate over whether there is suppression of unwanted memories, which

seems to face a paradox similar to the one facing self-deception, has gone on for nearl" a

centur"! It has, however, recentl" been put to rest b" Anderson et al(/BB8) who

produced not onl" evidence that there is such suppression, but also f95I evidence

concerning the neurobiolog" underl"ing it!

9emor" suppression re;uires people to override or stop the retrieval process! &ateral

prefrontal cortex is involved in stopping prepotent motor responses, switching tas# sets,

and overcoming interference in a range of cognitive tas#s! It ma" be h"pothesized,therefore, that people suppress unwanted memories b" recruiting lateral prefrontal cortex

to disengage hippocampal processing! (p! /1/)

*uppression is not tantamount to self-deception, since it can have the useful independent

function of minimizing distraction when we must handle a situation to which the

suppressed memor" does not pertain! 9emor" suppression can thus facilitate rational

thought, whenever the item suppressed needs to be brac#eted in order for the agent to

reason properl"! ut memor" suppression can facilitate self-deception! *uppression ma"

be used deliberatel" in willful self-deception to undermine memor" of evidence contrar"

to the desired belief! ut it ma" also figure into wishful self-deception without there

being an" intention to self-deceive! Kuite simpl", the discomfort engendered b" certain

memories can ma#e us wish to suppress them, which we do! Ghen that discomfort is

Finite Rational Self-Deceivers - 0,


18/30


19/30

about a persons intelligence is often mixed enough, even if its heavil" weighted to one

side, to enable the #ind of selective attending involved in the essential complex!

Do the other two functions of intentions, trac#ing and screening, also generate

desires li#el" to be implicated in self-deception+

Its less li#el" that the trac#ing function fre;uentl" brings about desires ripe for

self-deception! $he main purpose of trac#ing is accurac" and #nowledge, so its subsidiar"

desires wont be for states of affairs as much as information! Fven if the goal of the

intention is not well-defined, such as sending a child to a =prestigious> universit", the

trac#ing of related facts, such as admissions standards and tuition costs, will focus on

concrete things about which self-deception occurs onl" rarel"/?! I also dont see that the

screening function of intentions is a ma0or source of desires that produce self-deception,

since it will mainl" be the source of desires notto do particular things!


20/30

Aronson recalls an otherwise affable ietnam veteran he met who believed the

ietnamese werent human! Hlearl", the balance of evidence he has is on the side of the

belief that the" arehuman6 this explains wh" he felt guilt" at first about the atrocities he

committed! ut because of dissonance he desired that it not be the case that the beings he

#illed reall" were human6 this desire is the deceptive element in his self-deception that

the" werent! $o us, its plain that the civilians he #illed were human6 it would have been

plain to him too had it not been for self-deception! ut the point is that the ;uestion of

whether or not a certain entit" counts as humanis sufficientl" metaph"sical and

sufficientl" debatable, as weve seen from debates on abortion, that it falls into the sphere

of possible self-deception! Hases li#e that of the veteran can be multiplied!

L L L

efore moving to m" discussion of a paradigm case, let me re-cap the argument

thus far! $he challenge was to explain wh" humans have the capacit" for self-deception!

$he steps I too# to meeting this challenge were as follows! /E() I explained what features

of mind are involved in the capacit" for self-deception! (/) I argued7concurrentl" with

sa"ing what the" are7that those features are rational! $he b"product view follows: (1)

the capacit" for self-deception is a b"product of rational features! nce weve come this

far, two arguments will get us to favoring the evolutionar" view that the capacit" for self-

deception is a spandrel over the view that it is an adaptation! (8) .olding it is a spandrel

is perfectl" consistent with theprima facieplausible view that #nowledge enhances

fitness! () .olding that it is a spandrel is more parsimonious, for we dont need to posit

an additional adaptive function for self-deception!



21/30

2pplication of te Spandrel View to a Paradi&m ase

9" focus here has been to explain wh" and how wishful self-deception in

particular exists (as opposed to willful or dreadful self-deception/)! $he answer to the

;uestion of wh" we have the capacit" for self-deception is that the capacit" is a

b"product of mental features, each with useful functions, which in combination "ield

self-deception! $o see how the capacit" wor#s, lets consider a paradigm case!

$he case of wishful self-deception that I gave in section was that of the college

dropout! $he case I discuss here is more tragic, but also common! 9" focus is abuse in

familial or spousal relationships! Mnderstanding the ps"chological causes of the behavior

of the abuser is utterl" important, but we should also as# about the ps"cholog" of the

abused! Does self-deception ever pla" a role in #eeping a person in a relationship in

which he or she is abused+ I #now personall" of two such cases6 I dont thin# the"re

unusual! In one case, a woman abused b" her bo"friend admitted, after the relationship

ended, that she was self-deceived that the abuse would end! In another case, a man

explained, "ears into middle age, that he had been self-deceived up until age / that his

mother was a good parent, despite an ongoing pattern of abuse when he was a child!

$here is no reason to doubt these peoples accounts! .ow does self-deception arise in

such cases+ Ill present m" understanding of the case of the abused romantic partner!

$his person has man" cognitions7experience of abuse, #nowledge of patterns of

abuse7that on her own epistemic normsshould"ield the belief that her bo"friend will

continue to hurt her if she sta"s in the relationship! ut she has an intention(D*) to have

a health" relationship6 this intention has the subsidiar" desire that the abuse stop even

with the relationship continuing! $his desire is the deceptive element! Although she has



22/30

compelling evidence for the belief that the abuse will continue, she also has some limited

evidence (in the form of repeated promises from the bo"friend) that it will stop!

ecause of her desire for the abuse to stop and the relationship to be health", she

feels the sting of disappointment (FH) when attending to evidence suggesting abuse will

continue and comfort when considering memories of promises that suggest it wont (FH/

and FH?)! $he general inclinations to avoid discomfort and find pleasure (FH1 and FHE)

cause her attention (FH/) to be directed to the comforting evidence! Attention to this

evidence, b" FH, completes the self-deception! *he believes the abuse will stop even

with the relationship continuing!

$his process is aided b" the inertia of the web of belief (), because she started

out the relationship believing there would be no abuse and having man" other beliefs

about the goodness of her bo"friend supporting this belief! *ome of her memories of

abuse, furthermore, become suppressed (8)! *he greets her friends who tell her to get

out of the relationship with a higher degree of s#epticism (1) than she usuall" has!

ecause of her desire for the relationship to wor#, thoughts of relational harmon"

repeatedl" occur to her (/), ma#ing it easier for her to hold to the product of self-

deception! $he cognitive dissonance (D*/) arising from the tension between her

conception of herself as a self-respecting person and her behavior of sta"ing in an abusive

relationship adds to her desire that the abuse stop in a wa" that further facilitates self-

deception!

$his is how, b" a spandrel of the mind, an intelligent woman with ample evidence

of her bo"friends abusive character becomes self-deceived! $here is no need to posit an

intention to self-deceive or an adaptation thats designed for self-deception!



23/30

Self-Deception, Spandrels, and 3etodolo&y in 'volutionary Psycolo&y

I hold that the theor" of self-deception presented here has significant implications

for the field of evolutionar" ps"cholog"!

Its no secret that the dominant paradigm in evolutionar" ps"cholog" is

adaptationist! Honsider this passage from $oob" and Hosmides (@@):

ur cognitive architecture resembles a confederation of hundreds or thousands of

functionall" dedicated computers (often called modules) designed to solve adaptive

problems endemic to our hunter-gatherer ancestors! Fach of these devices has its own

agenda and imposes its own exotic organization on different fragments of the world! (p!

xiv)

If this is representative, then the program of evolutionar" ps"cholog" has been to explain

basic mental phenomena as features that were advantageous in the ancestral environment!

Githin this paradigm, 5obert $rivers (/BBB) attempts to explain self-deception as an

adaptation to help other-deception! I criticize $rivers theor" extensivel" elsewhere/@on

grounds of internal difficulties! ut m" main criticism of the dominant paradigm in

evolutionar" ps"cholog" at present7which also applies to $rivers7is that there is not

enough serious consideration of alternative evolutionar" h"potheses to adaptation! $here

are some examples of b"product views within evolutionar" ps"cholog"6 art, fiction,

religion, and music have all been claimed to be b"products!1B( on m" understanding, denotes

a b"product of phenot"pic traits!) ut those b"product claims tend to be about cultural

phenomena restricted to specific content domains and to not basic, highl" general mental

processes li#e self-deception! $he structural complexit" of the human mind should

prompt us to consider the possibilit" that man" more prominent features of mind than

have been recognized so far are structural b"products7spandrels!1I hope that this essa"



24/30

not onl" explains the capacit" for self-deception in particular as a spandrel, but also

provides an example of how one might construct a spandrel theor" in general!

$here are man" critics of evolutionar" ps"cholog"! ut there is a fallacious

implicit assumption that man" of them seem to share with those the" criticize: either a

prominent mental tendenc" is an adaptation or it is the result of environmental influence!

$his is no doubt simplif"ing the dialectic, but one gets the impression all traits are either

li#e teeth (adaptations) or li#e hairst"les (environmental)! $he battle is then fought over

which traits fall into which categor", with evolutionar" ps"chologists advocating

adaptations, their critics advocating environment, and moderates advocating a =complex

interpla"> of the two! In this dialectic, spandrels are often ignored! Ghen the possibilit"

of spandrels does come up, rarel" is a s"stematic theor" developed of how the spandrel in

;uestion actuall" wor#s!1/

I suspect the problem might be that no compelling general approach to

constructing ps"chological spandrel theories has been exemplified! In an Appendix on

evolutionar" ps"cholog" that summarizes the views of evolutionar" ps"chologists 5obert

Gright and 5ichard Grangham, a 5A (p! ?@)! &ac#ing a good

example of a mental spandrel, it seems, man" evolutionar" ps"chologists have felt free to

pursue the adaptationist paradigm unencumbered! urthermore, a close examination of

this Appendix reveals an important and curious presupposition of the authors! %oulds

arguments about spandrels are presented as criticisms of =evolutionar" ps"cholog",>

Finite Rational Self-Deceivers - 2"


25/30

without appending an" ;ualifier to the phrase =evolutionar" ps"cholog"> such as

=adaptationist!> $his shows that the authors implicitl" e;uate evolutionar" ps"cholog"

with adaptationist evolutionar" ps"cholog"! I thin# there is no good reason for this!

Mnderstanding the evolution of the mind'brain should involve both spandrel and

adaptationist explanations7and probabl" man" other #inds as well!


26/30

capacities for rational logico-mathematical and evaluative 0udgment, but we need to

cognize and survive with finite powers in the limits of finite time! If we had infinite

rational processing that allowed total ob0ective formation of 0udgments of good and bad

and true and false, then we wouldnt self-deceive! In the theoretical ideal of a perfectl"

rational angel, such 0udgments form the basis for all actions! In animals, finite perception,

memor", and processing necessitate a much more immediate cause of action in the

context of hostile environments: desires!

Desires alone, however, dont give rise to self-deception! Self-deception comes

from rationality in the context of a finite desiring mind! As Ive stressed, man" of the

features of which self-deception is a b"product are rational! $he abilit" to attend

selectivel" to evidence is rational6 it allows us to sift through the wash of inputs and

attend to the relevant ones! $he formation of beliefs on the basis of evidence is rational!

$he inertia of the web of beliefs is rational6 it prevents perpetual cognitive flux and ma"

be seen as characteristic of finite rational cognition as such! $he abilit" to appl" differing

degrees of s#epticism is rational6 it can be used to help us screen out falsehood! And the

abilit" to form complex intentions is rational6 it facilitates theoretical and practical

achievements! ut all of these rational abilities are implicated in the pervasive form of

human irrationalit" called self-deception!

$hus the capacit" for self-deception is not an incidental addition to our cognitive

ma#e-up! It comes from rationalit" in the context of finitude, two features thought central

to being human since the ancient %ree#s! $o summarize, our finitude contributes to our

capacit" for self-deception in two wa"s: (i) it necessitates desires for the sa#e of survival6

Finite Rational Self-Deceivers - 2&


27/30

(ii) it necessitates several formsselectivity in cognitive processing! Ghen the desires (i)

influence the selectivit" (ii), one is on the road to self-deception!

.umans are finite rational self-deceivers! Does this mean self-deception is

inevitable for humans+ I dont #now either wa", but I do thin# the propensityto self-

deceive is inevitable! *elf-deception can, however, be better avoided b" cultivating

cognitive habits that neutralize the aspects of mind that give rise to self-deception! ne

can confront discomforting evidence and accept it for what it is! I have also argued that

there are #inds of desires that humans characteristicall" have that are more li#el" to

engender self-deception: moral desires, desires for dispositional states of mind to obtain,

and desires for intangibles that are felt to be important! ne can be aware of these! A

person with such desires and mixed evidence is in a context ripe for self-deception!11

'ndnotes

Finite Rational Self-Deceivers - 2,


28/30

$here are actuall" different formulations of the paradox! $he one I give here presents self-deception as apparentl"

involving a straightforward contradiction: belief and not belief! ut, on a wea#er formulation, attributing self-deception

involves attributing a belief thatpand a belief that $p, which seems to involve positing something ps"chologicall" absurdwithout directl" resulting in logical contradiction on the part of the attributer! or either formulation the strateg" for

resolution will be the same: give a definition that captures the phenomenon without positing something absurd!/*ee Davidson (@/, @, @@)!1$here are not onl" these three puzzles! $hese are the most visible tip of the philosophical iceberg of problems surrounding

self-deception! *elf-deception raises ;uestions about: agenc", moral responsibilit", self-#nowledge, the role of emotion in

cognition, the distinction between what is perceived and what is inferred, what beliefs are in general, doxastic voluntarism,the distinction between tacit and explicit beliefs, belief attribution, and the value of truth!8Its important to note that this is a wor#ing definition! Im sure possible situations could be imagined in which thisdefinition is satisfied b" an irrational capacit" and a deviant causal chain, but this should do for now! Also, the word

=directl"> in clause (c) is important, since, as well see, rational capacities are indirectl"7via self-deception7implicated in

decrease in truth in the belief set!Gh" cant other #inds of emotional pollution besides desires muddle the epistemic process in a wa" that counts as self-

deception+ 5eflection suggests that self-deception ma" be caused b" fear, 0ealous", love, hate, shame, and even values and

ideolog"! 9" response in defense of m" definition will not be to exclude these, but to include them insofar as the" have a

motivational aspect! An aimless fear or paranoia cannot be the deceptive element in self-deception, although the" can cause

other cognitive foibles! ut love or fear can be if the" are directed in certain wa"s and motivates!?*ee note !EIve recentl" become aware of empirical data that supports the belief view! 5amachandran (@@) discusses an experiment

he does in which a patient in denial about her left side paral"sis grasps a tra" on one side with her right hand instead of inthe middle, suggesting she believesthat she could use her left hand to pic# up the other side! I also argue against the avowal

view in other wor#!Ghen I sa" that an agent has evidence thatp, I mean that she has cognitions that b" her own epistemic norms constitute

evidence that favors believingp!@$here are several other definitions of self-deception on offer in the philosophical literature! $he one closest to m" own is

9eles (/BB)! *ee also 5obert Audi (@), Jent ach (@), and %eorge %raham (@?), to name a few!B$his naturall" raises the ;uestion of what it means to have an epistemic norm! $here are man" complications lur#ing thatdo not fit into the scope of this essa"! In general, however, I thin# that self-deceivers count as self-deceived because the"

believe contrar" to patterns of belief formation (i) that the" t"picall" follow and (ii) that are rationall" 0ustifiable! In other

wor#, I define a ver" wea# form of self-deception in which a person believes contrar" to norms that she does hold but that

are not rational in the first place! I suspect intuitions will diverge as to whether such a state is appropriatel" labeled =self-

deception!> I avoid these complications here, however, and focus on paradigm cases!

$he sentence to which this is an endnote raises an important point! ne cannot simpl" eliminate the m"ster" of theexistence of self-deception b" sa"ing that the human mind is not that rational6 for in self-deception one violates epistemic

norms that one actuall" has! iolating ones own norms, I hold, is more m"sterious than violating abstract cannons! $hatswh" the relativization of self-deception to the agents norms is so important! As far as I #now, I am the first to define self-

deception in a wa" that ma#es use of this #ind of relativization!/*ee Audi (@@E), who argues that capturing such a tension is critical to a good definition of self-deception! Audi criticizes

9ele for not re;uiring an"thing that entails the tension! 9" definition, I hold, is not susceptible to such criticism!19ele (/BB) refers to it as =twisted self-deception!> In labeling this =dreadful,> I mean to suggest that the person engaged

in this t"pe of thin#ing or self-deception is motivated b" some form of dread or fear! I am not suggesting that these t"pes of

irrationalit" are, from an outside perspective, in an" wa" worse than the other forms!8It ma" be that there are more #inds of self-deception than the three discussed here that one can categorize on the basis of

relation between content of motivational element and content of belief! It is m" impression that most cases of self-deception

will be classifiable as one of these three t"pes! ut the =appropriatel"> in the content clause of m" definition is meant toleave open the possibilit" of other t"pes&azar (@@@) and 9ele (/BB) agree that non-intentional forms of self-deception are the most common!?Image: The 'merican 3eritage Dictionary of the 4nglish 7anguage: ourth Fdition, /BBB!EAlfred 9ele (@@E, /BB) also has put forth developed views on the etiolog" of self-deception! In other wor#, I compare

our theories in detail, but a detailed comparison would be out of place in this essa"! *o Ill confine m"self to pointing out

the two most important differences! irst, 9ele does not address an"thing resembling the b"product or spandrel theses, so

m" wor# spea#s more directl" to the second two puzzles identified in the Introduction! In fact, he doesnt raise these two

puzzles at all! *econd, 9eles $& model will have difficult" accounting for cases of self-deception that the spandrel modelhandles well, i!e!, cases in which the self-deceived agent comes to have a belief, the holding of which can onl" be seen as

having high sub0ective costs!


29/30

Different desires have their own characteristic stings6 for some the sting is anno"ance, for others anxiet", and for still

others simple disappointment! I ta#e the features I identif" here to be normal aspects of the human mind: the" obtain for the

vast ma0orit" of adult people!@&eDoux (@@?) ma#es the point in the cases of fear and anxiet", for example, that its not 0ust the anxiet"-causing

stimulus that people see# to avoid, but the anxiet" itself!/B$he fact that evidence needs to be mixed in order for self-deception to happen is crucial for explaining a ver" important

feature of self-deception! *elf-deception ver" rarel" occurs concerning perceptual beliefs6 in occurs much more often

concerning what might be called intangibles! $he reason for this is that evidence for or against perceptual beliefs is mixed

far less often!/Ghen I use the word =function,> I am using it in the pre-theoretic sense without a commitment to the functions being

adaptations (although, as indicated, I thin# the" might be)!//$he ;uestion naturall" ariseswhythe modulation of attention should be sub0ect to the general inclination to avoid

discomfort! ne answer is that ps"chological discomfort of an" sort is ver" often distracting in a wa" that prevents

accomplishment of goals, so one ma" need to attend awa" from what causes discomfort to accomplish something else!/1$his point is also important for addressing the third puzzle! ne might argue that the fact of self-deception is a reductio of

Davidsons interpretive view! (Hf! 4ohnston (@, p! ): =5ational connections are not constitutive and exhaustive of the

mental!>) ut this inference would be too ;uic#, for the involvement of FH in self-deception shows that there is a rational

step in this (overall) irrational phenomenon!/8$hroughout this discussion, Ive been tal#ing about desires as the form of motivation that gives rise to self-deception! ne

might reasonabl" wonder to what extent m" account will generalize to cases in which emotions li#e anger, love, or 0ealous"

constitute the motivational element behind the self-deception! I hold there are three relevant possibilities! irst, desires are

often associated with emotions and caused b" them, so that in such cases the explanation I can give will simpl" have theadded clause that an emotion was the source of the deceptive element desire! or example, if I am 0ealous of a rival, I ma"

want him to be unworth"6 this wanting will then be the deceptive element in self-deception that he is unworth"! *econd,

often intense emotions can cloud 0udgment without intervening desires, causing one to have un0ustified beliefs! *ome might

loosel" call this self-deception, but it is far enough from paradigm cases of self-deception that such tal# stri#es me as

misleading! $he phenomenon of simpl" having worse 0udgment when one is angr", for example, is no doubt interesting, but

it is not appropriatel" labeled self-deception! $hird, however, it is at least a conceptual possibilit" that an emotion should

constitute the motivational'deceptive element in self-deception withoutand associated desire doing the causal wor# that Iposit here! $he important ;uestion, then, will be: to what extent does m" account generalize to those cases+ *uppose, for

example, .arr"s disgustwith 4ennifer for stealing in the past leads him to believe self-deceptivel" that her children are

morall" blameworth" also! 9atters are tric#" here, for this sounds li#e a somewhat borderline case of self-deception! ut I

do thin# that important aspects of m" account will generalize! *elective attention will still, in this case, be driven b" finding

some evidence comfortable to attend to and other evidence not, and what causes the difference in whats comfortable will

still be the motivational state, i!e!, the emotion! urthermore, the facilitators I discuss in following sections, li#e the inertiaof the web of belief, can all still come into pla"! I will have to leave a fuller discussion of this topic, however, to another

occasion!/f the sort Kuine (@1) tal#s about! 5amachandran (@@, ch! E) discusses clinical cases of denial in wa"s that suggest

the sort of view Im advocating!/?$his comment raises the following difficult philosophical issue: what is the difference between perceptual beliefs and

beliefs arrived at via inference+ $his paper is not the place to address this topic! ut it seems to be that self-deception does

not t"picall" occur for perceptual beliefs7although see 5amachandran (@@, ch! E) for some extreme cases!/E$hese steps are not meant to represent the order of presentation! 5ather this is one wa" of representing the logical

structure of the argument!/Fxtending this model to cases of willful and dreadful self-deception will involve examining how the #inds of desire

constituitivel" involved in thoset"pes can trigger the sorts of selective attention and other processes here identified! or

reasons of space, I have not explored such an extension in this paper!/@an &eeuwen (forthcoming)!1B*ee, for example, loom (/BB) and Cin#er (@@E)!1Caul loom has pointed out to me in email that, 0ust as a matter of logic, the adaptations there are (however man" this

happens to be) will have as a conse;uence man" more b"products6 furthermore, most mainstream evolutionar"

ps"chologists are well aware of this! Ghile this is true, the point doesnt exempt the dominant paradigm from the criticisms

I offer here6 it is one thing to observe that there must be spandrels and another thing to wor# out rigorous theories of how

the" might wor#! urthermore, the centralit" of the mental processes h"pothesized to be spandrels is also being raised here6

m" point is that central mental processes li#e self-deception that greatl" and generall" influence behavior have been largel"exempt from the #ind of theorizing offered in this paper!1/I am also not suggesting that =adaptation or environment or spandrel> is an exhaustive dis0unction6 there is drift and other

factors that ma" influence phenot"pe as well!


30/30

11Gor# on this paper has benefited from exchanges with Jenneth $a"lor, 4ohn Cerr", Jrista &awlor, Dagfinn Ollesdal,

Frica 5oeder, 4ohn %abrieli, Flliott *ober, 4oanna idducia, &anier Anderson, 9ichael riedman,

! 5ort" and ! C! 9c&aughlin, er#ele", Mniversit" of Halifornia Cress" @/-/B!

Audi, 5! (@@E) PHomments,P see 9ele (@@E)!

ach, J! (@) PAn Anal"sis of *elf-Deception,P.hilosophy and .henomenological Research%(1): 1-1E!

aron-Hohen, *! (@@)8indblindness9 'n 4ssay on 'utism and Theory of 8ind, Hambridge, 9A, 9I$ Cress!

loom, C! (/BB) PIs %od an accident+P'tlantic 8onthly+60(): B-/!

ratman, 9! F! (@E):ntentions* .lans* and .ractical Reason, Hambridge, .arvard Mniversit" Cress!Damasio, A! (@@8)Descartes; 4rror,

Documents

Amagire de Sine