Sources of maladaptive behavior in ‘normal’ organisms

Sources of maladaptive behavior in ‘normal’ organisms

Accepted Manuscript Title: Sources of Maladaptive Behavior in ‘Normal’ Organisms Authors: Ralph R. Miller, Cody W. Polack PII: DOI: Reference: S0376-...

453KB Sizes 0 Downloads 26 Views

Accepted Manuscript Title: Sources of Maladaptive Behavior in ‘Normal’ Organisms Authors: Ralph R. Miller, Cody W. Polack PII: DOI: Reference:

S0376-6357(17)30394-7 https://doi.org/10.1016/j.beproc.2017.12.017 BEPROC 3566

To appear in:

Behavioural Processes

Received date: Revised date: Accepted date:

25-8-2017 8-12-2017 18-12-2017

Please cite this article as: Miller, Ralph R., Polack, Cody W., Sources of Maladaptive Behavior in ‘Normal’ Organisms.Behavioural Processes https://doi.org/10.1016/j.beproc.2017.12.017 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Maladaptive Behavior - 1 Sources of Maladaptive Behavior in ‘Normal’ Organisms

SC RI PT

Running Head: MALADAPTIVE BEHAVIOR

Ralph R. Miller and Cody W. Polack State University of New York at Binghamton

Ralph R. Miller Department of Psychology State University of New York at Binghamton

N

U

Mailing Address:

(607) 777-2291

FAX:

(607) 777-4890

M

TEL:

A

Binghamton, NY 13902-6000, USA

[email protected]

Submitted:

August 2017; revised December 2017

TE

D

E-mail:

Author Notes: RRM and CWP are both in the Department of Psychology, State University of New

EP

York at Binghamton, Binghamton, NY 13902-6000, USA.

RRM can be contacted at

[email protected]. Preparation of this manuscript was supported by NIH grant MH 033881.

CC

We are grateful to Tori Pena and Anna Tsvetkov for their comments on a preliminary version of the

A

paper.

Maladaptive Behavior - 2

Highlights: Behavior is usually adaptive; maladaptive behavior seemingly challenges evolutionary and reinforcement principles.



Most maladaptive behaviors arise from four families of factors.



Insufficient and excessive transfer of training is the root of many maladaptive behaviors.



The success and failure of transfer is often understood in terms of associative interference.

SC RI PT



Abstract

A basic assumption of most researchers is that behavior is generally functional, and indeed, in most But in a number of cases, some behaviors of

U

instances the function is obvious.

N

neurophysiologically ‘normal’ organisms appear to be maladaptive. Considerable research has

A

been conducted to understand the basis of such behavior as well as how the frequency of such

M

behavior can be reduced. Here we provide a brief panoramic review of the major sources of maladaptive behavior in neurophysiologically ‘normal’ organisms: a) altered environmental

D

contingencies relative to those faced by ancestral generations in their environment of evolutionary

TE

adaptation, b) altered environmental contingencies within the lifespan of the animal, c) linked behaviors in which the dysfunctional behavior is a linked companion of a more valuable beneficial

EP

trait, and d) the labeling of some behaviors as ‘maladaptive’ when more careful examination finds Most of our attention is on the consequences of altered

CC

that they provide net benefit.

contingencies across and within a generation, with altered contingencies within a generation

A

constituting a form of associative interference. The central issue in these two cases can be framed in terms of insufficient or excessive transfer of training resulting in maladaptive behavior. We discuss the functional basis of successful and unsuccessful near transfer (i.e., stimulus and response generalization) and far transfer (including rule learning and abstraction). Keywords: suboptimal behavior; dysfunctional behavior; altered contingency; stimulus and response generalization; transfer of training; associative interference.

Maladaptive Behavior - 3

Sources of Maladaptive Behavior in ‘Normal’ Organisms

SC RI PT

1.0 Introduction. Evolutionary theory assumes that animals are adapted to propagate their genes through selective pressures on survival and reproduction across generations, whereas reinforcement theory focuses on how animals modify their behavior within a lifespan to maximize pleasurable sensation. These two frameworks converge in that hedonic value was ultimately shaped by prior natural selection for attraction to sensory input that is positively correlated with genetic propagation.

U

Consistent with these assumptions, most behaviors are seen to be near optimal in serving these

N

ends within an organism’s ecological niche. However, the appreciable frequency with which

A

suboptimal and even dysfunctional (hereafter collectively called maladaptive) behaviors are

M

observed challenges simplistic applications of evolutionary and reinforcement theories. However, more careful examination of so-called maladaptive behaviors finds that these theories do not fail.

D

Rather, application of evolutionary and reinforcement theories requires consideration not only of

TE

the immediate circumstances of the observed behaviors, but also of the relevant prior experiences

EP

of the individual animal in question and its ancestors. For example, altruistic behavior is often cited as an example of behavior that is maladaptive with respect to the organism in question, but an

CC

appreciation of kin altruism and reciprocal altruism, especially when viewed within the social structure of the organism (e.g., our hunter-gatherer ancestors lived in small, tight knit bands),

A

permits us to recognize the underlying functional nature of many altruistic acts at least in the context in which the predispositions for these altruistic acts evolved. Here we present a broad framework for viewing maladaptive behaviors in general, without a lot of detailed examples and with no claim of originality for any of the principles proposed. [For an excellent review of some more specific learning processes that contribute to suboptimal behavior, see Zentall, 2016.] If there is anything new here, it is our bringing together these different principles and connecting them

Maladaptive Behavior - 4 to seemingly unrelated constructs within animal cognition and associative learning. When humans make choices that yield maximal available utility as opposed to other available choices that would yield less ideal outcomes, it is often said that they are behaving

SC RI PT

rationally (Kahneman, 2013; Kahneman & Tversky, 1979). But when nonhumans make similarly optimal choices, many researchers are inclined to talk about the underlying ‘automatic’ behavioral principles at work without appealing to any sort of conscious reasoning as is implied by rationality. The tendency for researchers to provide different accounts of similarly adaptive or maladaptive behaviors seen in humans and nonhumans is likely fueled by our introspecting about our own

U

behaviors (see Polack, McConnell, & Miller, 2013, for a discussion of this issue with respect causal reasoning). Here we will regard so-called conscious reasoning as either (a) an emergent property

N

of underlying behavioral principles (i.e., a different level of analysis rather than an alternative

A

account), or (b) basically epiphenomenal with respect to behavioral choice. Without taking a

M

position with respect to these two alternatives, we will focus on the presumed underlying behavioral

D

principles as opposed to a higher-order cognitive account in terms of rationality.

TE

1.1 Defining maladaptive behavior in ‘normative’ organisms. Our concern is the basis of maladaptive behavior by organisms with neurohormonal

EP

systems that are normative. Thus, here we do not consider the numerous maladaptive behaviors

CC

that can and do arise from abnormalities in an individual organism’s neurophysiology relative to the norm for the species. We acknowledge that there is often not a clear distinction between these

A

two families of sources of maladaptive behavior. However, to first approximation, we differentiate between maladaptive behaviors that would be seen in the majority of a population if it were exposed to the precipitating environmental contingencies in question and maladaptive behaviors that would appear in only a small fraction of the population if it were exposed to the same precipitating environmental contingencies. Our present focus is exclusively on sources of the former class of maladaptive behaviors while recognizing that they may also contribute to the latter

Maladaptive Behavior - 5 class in certain cases. 2 Four sources of seemingly maladaptive behavior. Looking across a wide variety of maladaptive behaviors, we think that most if not all of them can be

SC RI PT

assigned to one of four categories: (a) those behaviors that arise due to a change in environmental contingencies across generations, (b) those that arise from a change in contingencies experienced by the organism in question, (c) those that are caused by the same mechanisms that also give rise to functional behaviors, and (d) those that are in fact not really maladaptive but their functional values have not been fully recognized. Below we briefly speak to each of these four sources of

U

so-called maladaptive behaviors. But our focus is primarily on the first two which are at the root of

N

how organisms deal with changing contingencies, first as a species, then as individuals.

A

2.1 Type a: Change in environmental contingencies across generations.

M

Environmental contingencies concerning which responses lead to optimal reproductive When the environment

D

success and hedonistic reinforcement are not constant over time.

TE

provides a new contingency, animals often learn within their lifetimes how to exploit the contingency (e.g., Krebs, Kacelnik, & Taylor, 1978). But when the new contingency has been in

EP

effect over numerous generations, genetically based predispositions favoring the occurrence of the behavior with minimal or no learning are favored by natural selection provided that such biases

CC

happen, largely by chance, to arise in the genome (Baldwin, 1896). Critically, behaviors that are so functional that they are acquired widely across a species are exactly those behaviors (or at least

A

the mechanisms favoring those behaviors) that will be selected for if and when predispositions favoring them happen to arise in the genome; moreover, that the behavior was widely learned changes the environment in which genome selection occurs to one more strongly favoring selection of the genetic predisposition (i.e., Baldwin Effect). Once the genetic predisposition for a specific behavior has been selected for, learning anew in each subsequent generation is no longer necessary, although in some cases across generations organisms evolve specialized

Maladaptive Behavior - 6 predispositions to learn rapidly ecologically important relationships within a generation (e.g., Garcia effect; Garcia, Hankins, & Rusiniak, 1974). In the most extreme cases of this, the predisposition is often called an ‘instinct,’ although there is surely a continuum from ‘unlearned’ behaviors such as breathing, pecking for food in the young herring gull (Tinbergen & Perdeck, 1951), swimming by

SC RI PT

salamander tadpoles (Carmichael, 1926), and imprinting (e.g., following behavior) by ducklings (Lorenz, 1935), to learning of behaviors acquired through reinforcement principles such as solving complex mathematical problems. Even instinctual behaviors need at least some limited degree of experience to be expressed or refined (Balsam & Silver, 1994), and complex behaviors are often built on more competencies that are dependent on genetic predispositions (e.g., math is an

U

extension of numerosity; Xu, Spelke, & Goddard, 2005). Behavior is never ‘merely’ instinctual or

N

purely learned, but emerges from an interaction of instinctual predispositions modified and

A

expanded upon by experience. The fecundity of abilities resulting from that interaction

M

subsequently impacts the genetic groundwork of future generations.

D

Once natural selection has created a genetic predisposition for a particular behavior, the animal is advantaged in requiring little or no learning for the functional response to appear in the

TE

animal’s behavioral repertoire. Ordinarily such predispositions are highly functional, resulting in

EP

differences in ease of learning different cue-to-consequence effects (Garcia et al., 1974) and cue-to response to-consequence effects (Foree & LoLordo, 1973). Such selective predispositions to

CC

learn certain relations extend from no specific experience (i.e., learning) at all being necessary for the behavior to be expressed to situations in which learning fails to occur even following numerous

A

training trials. However, if the environmental contingency subsequently changes, the genetic predisposition favoring or discouraging the behavior could now be maladaptive (reviewed in Seligman & Hager, 1972). Assuming that appropriate genes exist within the genome, natural selection will, over sufficient generations, reduce the predisposition for the maladaptive behavior and promote a predisposition for a more adaptive behavior. The rate of the selection against the maladaptive behavior depends on the magnitude of the dysfunction (i.e., degree of disadvantage in

Maladaptive Behavior - 7 terms of reproduction) that it produces given the population’s environment. Until such genes come to exist by chance and are selected for, the now maladaptive behavior can be viewed as an example of phylogenetic inertia. More generally, phylogenetic inertia is defined as the rate of natural selection, which is constrained by what prior selection has brought forward, being slower

SC RI PT

than the recent rate of environmental change.

Among the most widely cited examples of a change in the environment that makes a genetically-based behavioral predisposition maladaptive is dietary choice by humans (Armelagos, 2014; Lieberman, 2016; Woods & Begg, 2015).

Our hunter-gatherer ancestors regularly

U

experienced feast and famine depending on their recent success in obtain food. Killing a large game animal provided a huge amount of meat, more than could immediately be consumed.

N

Without refrigeration or canning, protecting the excess food that could not be immediately eaten

A

from other humans, other predatory species, and bacteria was an enormous challenge. Humans

M

(as well as other species) addressed this challenge by ingesting as much food as they could hold,

of food scarcity.

D

converting the immediately surplus energy into fat that could be drawn upon during the next period That is, the safest place to store immediately excess calories was as fat Consistently carrying large amounts of excess calories has negative

TE

underneath the skin.

EP

consequences for health. But in our ancestors’ environment of evolutionary adaptation (EEA), excess calories were only occasionally present to be ingested and the benefit of the protection from

CC

starvation in future times of scarcity outweighed any deleterious health consequences of intermittently carrying excess calories as fat.

In contrast, today most humans in developed

A

countries live with consistent, abundant supplies of food. The optimal (i.e., healthy) strategy under these circumstances would be to ingest only the number of calories immediately necessary to function. But our genetic predisposition to store ingested excess calories as fat under the skin is now maladaptive given the relative absence of widespread scarcity that had been prevalent in our EEA. One result of this predisposition is today’s obesity epidemic.

Maladaptive Behavior - 8 A laboratory-based example of the same principle is a phenomenon known as negative automaintenance (Williams & Williams, 1969). Autoshaping and automaintenance refer to the findings that if a species-typical response (for example, a feeding response such as pecking by a pigeon) is followed by food, the animal will increase its emission of that response (i.e., autoshaping)

SC RI PT

and then maintain the response (i.e., automaintenance) even when there is no causal contingency between the response and the food (Brown & Jenkins, 1968; Sidman & Fletcher, 1968). Autoshaping occurs rapidly even in the absence of a reinforcement contingency; alternatively stated, the animal has a strong predisposition to make feeding responses at cues that have been paired with food even when the responses are not required to obtain the food. The predisposition

U

has been selected for over generations due to the fact that in the animal’s (e.g., pigeon’s) natural

N

environment, autoshaping is highly functional because the autoshaped response usually does

A

increase the animal’s chances of obtaining food in the species’ natural habitat. Moreover, direct

M

prior experience of the response being followed by food by the individual animal in question may also contribute to the predisposition to autoshape quickly.

However, the specific associative

D

mechanism underlying autoshaping is relatively insensitive to the actual strength of the underlying

TE

instrumental contingency between the specific response and reward. Thus, the response is still acquired and retained even in those relatively few instances (often artificially created in the

EP

laboratory) in which there is no causal relationship. This predisposition to autoshape is present presumably because the benefit of rapid learning in those instances in which there is a causal

CC

relationship outweighs the cost of responding in those few instances in which there is not a causal relationship (Hearst & Jenkins, 1974; Locurto, Terrace, & Gibbon, 1981). That is, in foraging

A

behavior, a false alarm is typically less costly than a miss. Critical to the point being made here, negative automaintenance refers a situation (normally made in a laboratory), in which an experimenter creates a negative contingency between the response in question (e.g., a pigeon pecking at a keylight for which illumination has in the past been followed with food) and the delivery of food. Now, rather than responding simply not being necessary to obtain food as in autoshaping,

Maladaptive Behavior - 9 responding prevents the delivery of food which would otherwise have occurred). This sort of omission schedule does reduce responding to a degree, but it fails to stop responding entirely despite each response decreasing the delivery of food. The animal’s predisposition to relentlessly continue to respond over many trials even when responding reduces food delivery is a distinctly

SC RI PT

maladaptive behavior arising from the species’ genetic predisposition to make feeding responses that are sometimes paired with food, even when such responding actually reduces receipt of food. Autoshaping and automaintenance both can be disassembled into whether the animal responds toward the cue (i.e., sign tracking) for a reward or toward where the reward will

U

momentarily be presented (i.e., goal tracking; Anselme, Robinson, & Berridge, 2013; Chow, Smith, Wilson, Zentall, & Beckmann, 2017; Patitucci, Nelson, Dwyer, & Honey, 2016). Presumably goal

N

tracking is more functional because in naturalistic situations approaching the location of impending

A

reward decreases the delay to reward and/or increases the probability of reward. Sign tracking

M

presumably occurs at least in part because in animals’ ancestral ecological niche (and current

D

ecological niche) cues ordinarily occurred in spatiotemporal proximity to the rewards that they signal (Silva, Silva, & Pear, 1992). The maladaptive nature of sign tracking arises from unusual

TE

situations, often laboratory situations, in which a cue and its reward are spatially separated.

EP

Schedule-induced polydipsia (e.g., Falk, 1961) is another laboratory-based example of how a strong behavioral predisposition can become maladaptive when environmental contingencies

CC

change. Rats, and mammals in general, tend to drink water after eating. When rats are given very lean amounts of food at regular intervals, they will drink excessive amounts of water, which

A

puts them at risk of hyperhydration. In each of the above examples, we see that a genetic predisposition to emit a specific behavior, that was functional for past generations but is not today, is a source of maladaptive behavior. One might look at this as the environment providing ‘unnatural’ contingencies (feasts without famines and the imposition of an omission schedule where a reward schedule has

Maladaptive Behavior - 10 prevailed for many generations) that basic behavioral processes were not designed to address (see Zentall, 2016). Some additional examples of maladaptive behavior in humans behavior include loss aversion (i.e., animals give more weight to avoiding losses than obtaining gains; Kahneman & Tversky, 1989), cognitive dissonance (i.e., making a decision based on a prior

SC RI PT

decision even when consistency between the two decisions is maladaptive; Festinger, 1967), risk taking (i.e., choosing a higher risk course of action even when over sufficient trials it will not provide less gain than lower risk actions, e.g., gambling; Pawlowski, Atwal, & Dunbar, 2008), and sunk cost effects (i.e., continuing with a low reward behavior, when a higher reward behavior is available, as a result of having previously performed the low reward behavior; e.g., Arkes & Ayten, 1999). Each

U

of these effects is observed in nonhumans as well as humans and they can all be accounted for in

N

terms of predisposed mechanisms that are generally functional in the species’ ancestral (and often

A

current) ecological niche. In each case, a genetically (or experientially) predisposed associative

M

mechanism results in maladaptive behavior in a situation deviating from the contingencies that the

D

animal ordinarily encounters in its natural habitat or its ancestors’ natural habitat. To be sure, we are not suggesting that all maladaptive behaviors in this category were

TE

functional at one or another point in the animal’s own past or its ancestral history. Often the

EP

behavior did not even occur previously. Instead, it was the processes underlying the maladaptive behavior that yielded functional behavior in past generations. The rather different environmental

CC

conditions of long ago may have caused these processes to produce behaviors that were functional in the species’ EEA, but under modern environmental conditions a vastly different pattern of

A

behaviors manifest that are sometimes maladaptive. Our ancestors were good hunter-gathers who experienced benefits from having a strong

appetite for fats and sweets. One might ask why modern people often fail to select healthy fruit over a candy bar today?

The answer is that there were no candy bars 10,000 years ago.

Consequently, there was no benefit in evolving a limit to sugar intake because excessive sugar was

Maladaptive Behavior - 11 simply not part of our EEA.

Candy bars are a supernormal stimulus, in which the naturally

occurring eliciting stimuli that we evolved to respond to are enhanced to produce an exaggerated response (e.g., Goodwin, Browne, & Rockloff, 2015). In fact, eating all available food one could hold was beneficial due to the frequency of shortages of food. Today, excess sugar is part of our

SC RI PT

current ecological niche, but natural selection lags behind. Large scale availability of excess sugar is less than 100 years old, and the phylogenetic inertia of our taste preference has resulted in a deleterious mismatch between our EEA and the current availability of sweets (Lenoir, Serre, Cantin, & Ahmed, 2007). Admittedly, the construct of EEA must be qualified because it is not a single moment in the past, but a composite across a span of time in our ancestral past and the EEA

U

span for each trait likely differs.

N

A further example of phylogenetic inertia can be seen in the family of maladaptive behaviors

A

that arise from what is often called delayed discounting, which refers to choice behavior indicative

M

of a reward delayed into the future being given lesser value than a more immediate, smaller reward.

D

Speculatively, in our ancestral past, the future was much less certain than was the present, whereas in modern human society we have a greater degree of certainty concerning events in the

TE

future. For example, we need not store food at home because we know with relatively high Modern humans show

EP

certainty that tomorrow the supermarket will not be out of food.

maladaptively steep amounts of discounting for future rewards (e.g., Kirby & Marakovic, 1996).

CC

However, steep temporal discounting by our ancestors likely was functional in that in our EEA

A

rewards delayed were often rewards never received. Most changes in the environment leave organisms as worse fits to the current environment

than we would be if the contingencies had not changed, but some environmental changes occasionally improve our fit to the current ecological niche. For example, in recent centuries, people have come to spend less time outdoors, where sunlight is prevalent and the eye is forced to frequently change focus, and more time indoors, where lighting is dimmer and often focal distances

Maladaptive Behavior - 12 of the eye are held constant for long periods of time (e.g., Sivak, 2012). These factors have resulted in large numbers of people developing problems in focusing their eyes, which is maladaptive. However, the mass production of eyeglasses and development of corrective eye surgery both constitute relatively recent changes in our environment that dramatically reduce the

SC RI PT

penalty for weak eyesight. Inherent to this view is that natural selection is ordinarily slow, often taking many generations for genetically-based predispositions favoring maladaptive traits to be eliminated from the gene pool, while predispositions favoring traits that are merely not functional rather than maladaptive are never selected against. Moreover, the rate of the selection against maladaptive traits depends strongly on the magnitude of the dysfunction (i.e., the resulting impact Many of humans’ modern health

U

on fitness) given the current environmental contingencies.

N

concerns arise from the fact that our life expectancy has increased to engender previously unseen

A

costs associated with our behavioral predispositions. Moreover, it is quite possible these negative

M

behavioral predispositions will persist until they begin to directly impinge on reproduction.

D

2.2 Type b: Change in environmental contingencies within a life span.

TE

In contrast with maladaptive behaviors that arise from changes in environmental contingencies which render maladaptive some genetically-based behavioral predispositions that

EP

were beneficial to one’s ancestors, other maladaptive behaviors can arise from changes in environmental contingencies that occur within the life span of a single organism. This is more apt

CC

to occur with behaviors that are relatively plastic but, once acquired, minimize feedback to the organism concerning the new contingency.

One might think of such situations as

A

experientially-based behavioral predispositions, in contrast with the previously described genetically-based behavioral predispositions. A classic example of this situation is Gwinn’s (1949) so-called vicious circle behavior. Gwinn studied the behavior of rats in a circular runway that had a grid floor. During Phase 1, one group of rats was trained to escape a footshock which was delivered in both the start compartment

Maladaptive Behavior - 13 and the circular runway (but which was absent when the rats completed the circle) by running around the runway, thereby reaching the start compartment from which they had begun. This task was mastered quickly, with the rats learning to initiate running as soon as the door between the start compartment and rest of the runway was opened. During Phase 2, the same procedure was

SC RI PT

used except no footshock was delivered in the start compartment. Optimal behavior during Phase 2 was to simply stay in the start compartment which was now free of footshock, rather than run around the runway which was still electrified except for the start compartment. Rats that had not been exposed to Phase 1 training learned to stay in the start compartment within a few trials, often only one. In contrast, the rats that had been given Phase 1 training continued over a large number

U

of trials to run around the runway, thereby receiving footshocks that they could have easily avoided

N

by simply staying in the start compartment. Critically, Phase 1 training left the rats so quick in

A

exiting the start compartment on their trip around the runaway that they seemingly failed to perceive

M

the new contingency that was in force during Phase 2, that is, the absence of footshock in the start compartment. Alternatively stated, one might view such maladaptive behavior as an example of

D

excessive transfer of training from the Phase 1 contingency to the Phase 2 task (for elaboration,

TE

see Brown, 1969; Melvin, 1971).

Such excessive transfer is increased over what it might

otherwise be if the behavior acquired during Phase 1 impairs the organism’s ability to learn the new

EP

contingency introduced during Phase 2. For example, in an active avoidance paradigm, the avoidance behavior is well maintained even though the reinforcer is no longer present (e.g.,

CC

Sidman 1953). A skilled subject would fail to detect a change in the environment that somehow removed the aversive event altogether, thereby allowing the now unnecessary avoidance behavior

A

to persist. A clinical example of this is the persistence of human neurotic behavior in which person’s avoidance behavior prevents the person from experiencing the absence of the aversive event. Specifically, in cases of PTSD, avoidance and suppression of the aversive memory is a common maladaptive strategy linked with persistence of intrusive thoughts and memories (e.g., Davis & Clark, 1998).

Maladaptive Behavior - 14 It is important to distinguish maladaptive behaviors that arise from excessive transfer of nontarget training to the test of target training as discussed above, which can viewed as a form of associative interference, from purely insufficient transfer of target training from the target training experience to testing of target training. The latter situation is possible only in situations in which

SC RI PT

the test situation differs to some degree from that of target training. In such cases, insufficient transfer of training amounts to the retrieval cues at test being inadequate to activate the target memory, which can be viewed simply as generalization decrement going from target training to target testing. In this case, there is not necessarily any interfering memory.

U

In contrast, maladaptive behavior arising from changing environmental contingencies experienced by a single organism (i.e., within a generation) often constitutes a form of interference,

N

usually proactive interference. That is, this situation can be viewed as the test trial’s retrieval cues

A

activating the memory of a previously-acquired, but currently irrelevant contingency. The result of

M

this irrelevant memory activation could be either the emission of behavior that is inconsistent with

D

the emission of currently functional behavior (i.e., response competition between the currently functional response and the response that was previously functional in the presence of the current

TE

retrieval cues) or the inhibition of retrieval of the currently functional association (i.e., associative

EP

interference) as a result of activation at test of the interfering association. The more general underlying assumption here is that activation of a memory inhibits activation of other memories that

CC

share some components but not other components (i.e., common cues but different outcomes;

A

Bouton 1993; Miller & Matute, 1998). We often think of a failure of transfer of training as maladaptive, which it is provided that the

training and test situations include the same environmental contingencies. But when the training and test contingencies differ, as in the vicious circle paradigm described above, transfer of training is maladaptive. That is, transfer of training from one situation to another is adaptive depending on the similarity of contingencies between the two situations.

However, for humans the more

Maladaptive Behavior - 15 common problem is insufficient transfer rather than excessive transfer of training, as exemplified by the challenge of getting students to transfer principles (e.g., rule learning) that they have learned across diverse situations. But there are those situations in which transfer is maladaptive because different contingencies prevail between the environment of training and that of testing. Digressing,

SC RI PT

one might ask why undergeneralization is more often a problem for humans than is overgeneralization. We speculate that in ancestral times reinforcement contingencies were not as consistent across contexts as they are in modern times.

Nevertheless, maladaptive

overgeneralization does sometimes occur (e.g., post-traumatic stress disorder).

U

2.3 Type c: Linked behaviors.

N

Some maladaptive traits, including behaviors, are not selected against by natural selection

A

or reinforcement principles because the genetically- or experientially-based predispositions

M

favoring the maladaptive traits arise from proximate mechanisms that also favor adaptive traits. An early identified example of a maladaptive trait arising from a genetically-based predisposition is

D

sickle-cell anemia, a condition in which the red blood cells have an abnormal shape that reduces

TE

their ability to carry oxygen through the circulatory system. The malady causes victims to become exhausted after relatively little exertion, and in extreme cases can be fatal. One might think

EP

natural selection would eliminate the predisposition from the gene pool of a breeding population. But despite many generations over which selection might have been expected to occur, sickle-cell

CC

anemia in still prevalent in many parts of sub-Sahara Africa. Subsequent research found that the predisposition for sickle-cell anemia is caused by a single gene on each member of a single pair of

A

chromosomes. Critically, the same gene provides protection against malaria, a virulent scourge of much of Africa south of the Sahara Desert. Moreover, this protection against malaria is provided by even one of the two paired chromosomes having the relevant allele; that is, the protection against malaria is a dominant trait. In contrast, the sickle-cell anemia is a largely recessive trait in that under most environmental conditions it occurs only when both of the paired chromosomes

Maladaptive Behavior - 16 have the allele (Serjeant, 2010). As individuals heterogeneous for the gene locus are necessarily more numerous than individuals having two copies of the gene, natural selection favors retaining the gene in the population both for its protective consequences and because the number of protected people necessarily exceeds those afflicted with sickle-cell anemia. This explains why

SC RI PT

natural selection has not reduced the prevalence of the gene favoring the maladaptive trait of sickle-cell anemia in populations living in regions in which malaria is pervasive. Natural selection over generations seemingly has prevented the relevant gene from becoming frequent in populations living in areas where malaria is not a problem.

U

The relationship of two (or more) traits being modulated by the same gene, as in the case of sickle-cell anemia is referred to as pleiotropy. However, our point concerning predispositions for

N

multiple traits, some maladaptive and others functional arising from the same genes, need not be

A

restricted to so-called single-gene traits. The traits could be polygenetic traits modulated by an

M

array of genes so long as there is a large degree of overlap between the genes relevant for the two

D

traits. Linked traits refer to traits that are linked by common genes, but are unrelated at the level of phenotype, and differ from the previously mentioned changes in contingencies across generations

EP

ecological niche.

TE

which concern a change in the value of a single trait as a function of modification of the species’

In the prior example of a single gene (or array of genes) that produces a predisposition for a

CC

beneficial trait also preventing elimination of a predisposition for a maladaptive trait, the maladaptive trait was not a specific behavior but the general dysfunction of muscles that

A

accompanies anemia. As Darwin (1871) noted, the same rules of natural selection apply to behavioral traits as apply to anatomical and physiological traits because behavior is the product of neuroanatomy and neurophysiology. Thus, it is not surprising that there are numerous examples in which the maladaptive trait in a pair of linked traits has behavioral specificity. As discussed in section 2.1, humans have evolved an appreciable taste for calorically

Maladaptive Behavior - 17 dense foods that has resulted in a number of negative health consequences due to a mismatch between our EEA and the modern availability of foods. Moreover, the flexibility with which we can learn an association between a flavor cue and high caloric density has also become harmful (Swithers & Davidson, 2008) in that such flavor learning undermines the regulatory role of cues we

SC RI PT

would instinctually use to signal satiety. Our capacity to learn such relationships within a lifetime leads to maladaptive ingestive behavior as a result of the very same basic learning mechanism that usually serves us so well.

Our sympathetic nervous system developed in order to aid humans in facing diverse

U

environmental threats that plagued us, particularly in our EEA. An externality of changes in sympathetic activation is the pleasurable sensations that accompany both increases and

N

decreases in sympathetic arousal. The result is that many humans seem to actually enjoy taking

A

risks to induce such changes. Consequently, some individuals are driven to pursue dangerous

M

situations, not necessarily for the potential gains that might result from a successful hunt, but

D

merely for the adrenaline rush. Risk-taking in general is a potentially adaptive trait, as risky decisions sometimes lead to highly beneficial outcomes. But we frequently see individuals who

TE

seek dangerous thrills as a form of entertainment with no further accompanying benefit. What is

EP

adaptive about paying money to willingly jump out of a perfectly good airplane? One might expect natural selection to reduce this degree of risk seeking, but the benefits of sympathetic arousal in

CC

threatening situations even in modern society will likely prevent the predisposition toward risk

A

taking from being weeded out of the gene pool. That some species of moles are blind might be viewed as another instance of a maladaptive

trait being linked to a beneficial one. Moles live almost entirely underground, so vision would be of limited value to them; however, it would still be of some value. But some of the parts of mole brain that would otherwise be devoted to vision now serve tactile sensation, which is highly useful to animals living underground. Thus, although blindness in moles can be regarded as maladaptive, it

Maladaptive Behavior - 18 seems linked to their having heightened tactile acuity which provides more benefit than the cost of blindness. 2.4 Type d: ‘Maladaptive’ behaviors which are not really dysfunctional.

SC RI PT

A fourth category of so-called maladaptive behaviors are those that are not really maladaptive; rather their consequences are functional although their benefits are not widely recognized.

Most behaviors have both merits and demerits in the context of a particular

environment (i.e., set of contingencies). However, we often fail to see the merits, leaving us thinking that a given behavior is net dysfunctional when it is not. There is some degree of overlap

U

here with maladaptive behaviors that are linked with adaptive behaviors; however, here we are not

N

speaking about linked traits, but rather a single trait that may have both detrimental and beneficial

A

effects, with the beneficial effects being under recognized. For example, after losing a battle with

M

a male conspecific, males of most species experience a decrease in serotonin and testosterone, which results in depression-like behavior (La Frenière, 2010). This depression-like behavior is

D

often viewed as maladaptive. But it is now generally recognized as part of a behavioral syndrome

TE

that reduces the likelihood of the animal becoming engaged in another battle before it has fully healed from the first, which is usually functional. Another behavioral example is the cries of a rat

EP

pup outside the nest that were once viewed as maladaptive in that they could attract predators. However, the cries were subsequently seen to attract the dam to retrieve the pup back to the

CC

warmth of the nest (de Ghett, 1978).

A

An example of a seemingly maladaptive behavior that upon reflection can be seen to be

beneficial is that in some species (e.g., lions), when the male within a breeding couple is displaced by a new male, the normally protective female makes little or no effort to protect her young while the new male kills them (Bertram, 1975; cf. Packer & Pusey, 1983). This lack of direct action to protect the offspring by the female which typically is defensive of her young was initially viewed by observers as maladaptive. But further consideration made clear that her behavior was in fact

Maladaptive Behavior - 19 beneficial to her because starting a new family with the incoming male would obtain his cooperation in raising the new young. Another example of a superficially maladaptive behavior is fainting (i.e., syncope), which is

SC RI PT

unconsciousness arising from insufficient cerebral blood circulation during a temporary failure of the circulatory system. Fainting in humans was once viewed as maladaptive behavior due to the danger of falling and remaining incapacitated. However, more recent consideration suggests that a fainting person in becoming prone improves blood flow to the brain (Gert van Dijk, 2003).

A general cognitive example of a seemingly maladaptive trait that is actually adaptive in

U

most situations is seen in one source of day-to-day forgetting of acquired information. Almost

N

everyone wishes that they were less prone to forgetting. However, owing to the limited size of

A

working memory, forgetting of currently irrelevant information often reduces interference with

M

immediately relevant information (e.g., Luria, 1968). The forgetting of irrelevant information in favor of recall of immediately relevant information is readily demonstrated in the phenomenon of

D

retrieval-induced forgetting (RIF; Anderson, Bjork, & Bjork, 2000; Soares, Polack, & Miller, 2016).

TE

In a RIF situation, subjects in Phase 1 are asked to learn two independent sets of categories, each with exemplars (e.g., fruit-pear, fruit-banana, clothing-shirt, clothing-hat). Then in Phase 2, they

EP

are asked to recall some of the exemplars from one of the categories (e.g., recall pear when presented with fruit-p___). On a final recall test of all items, exemplars that were not recalled but

CC

were related to those that were recalled in Phase 2 (e.g., banana) fail to be recalled as readily as exemplars that were initially studied but unrelated to the exemplars sought in Phase 2 (e.g., shirt

A

and hat). Seemingly, access to the fruit-banana association was inhibited during Phase 2 in order to facilitate activation of the fruit-pear association and this inhibition carried over to the final recall test. Thus, the mechanisms responsible for forgetting are seen to have a net beneficial value in Phase 2. Although we are intensely aware of the consequences of forgetting on those relatively few occasions that we cannot recall something we previously learned, we never think to be grateful

Maladaptive Behavior - 20 for failing to recall information that would interfere with our performance on a current task (Bjork, 1989). 3 Bases of generalization and transfer.

SC RI PT

In section 2, we asserted that considerable maladaptive behavior arises from excessive, or sometimes insufficient, generalization and transfer of training. Here we consider the nature and foundation of generalization and transfer, and how they each sometimes lead to maladaptive behavior.

U

3.1 Stimulus generalization.

Input from the

N

No stimulus complex ever reoccurs exactly as it initially occurred.

A

environment inevitably differs, at least slightly, from trial to trial due to variation in environmental conditions. Additionally, the internal milieu of the organism (e.g., affective state) is in constant flux,

M

adding to the variance in a perceived stimulus between trials. If organisms used previously

D

acquired information to influence their behavior only when the present cues were perceived to be

TE

exactly the same as the cues on a prior learning trial, prior learning would never influence subsequent behavior, thereby rendering learning useless.

Stimulus generalization refers to

EP

behavioral control by cues at test that differ to a degree from those of training. As learning without stimulus generalization would be useless, it is reasonable to think that the capacity for stimulus

CC

generalization co-evolved with the capacity to learn. Whereas some degree of generalization is essential for learning to be useful, unlimited

A

stimulus generalization would also poorly serve animals, as animals need to respond differentially depending on the current environmental conditions.

Generally speaking, as the differences

between test cues and training cues grow greater, the cue-outcome contingencies are less apt to remain similar to those of training. Hence, unlimited stimulus generalization would have animals responding based on contingencies that were relevant during training, but no longer hold at test.

Maladaptive Behavior - 21 Such overgeneralizing would lead to maladaptive behavior (Reiss, 1980). Prime examples of overgeneralization leading to maladaptive behavior are anxiety disorders such as PTSD and specific phobias. Fear in these situations is not to just any arbitrary set of test cues, but to cues similar to cues for which there was a functional basis for fear in the animal’s evolutionary or

SC RI PT

experiential past. Sudden onset, loud noises trigger maladaptive fear responses in ex-soldiers because of a similarity of these noises to noises previously heard in the battlefield. Phobias tend to be overgeneralized fear to innocuous objects that in modified form are things that animals (e.g., people) have a basis for fearing such as large carnivores, poisonous spiders and snakes, and great heights. People rarely develop phobias to flowers or rabbits (e.g., Cook & Mineka, 1989; Öhman,

U

& Mineka, 2001; Seligman, 1971; cf. Watson & Rayner, 1920). Conversely, people often exhibit

N

inadequate fear of things that they should fear such as guns and cars, perhaps because they had

A

no counterpart in ancestral times. More generally, the benefit of stimulus generalization when

M

training and test cues are similar and the reduced benefit of stimulus generalization when training and test cues are relatively dissimilar anticipate the observed decrease in stimulus generalization

D

as the training and test cues become more dissimilar. This accounts for increasing stimulus

TE

generalization decrement as training and test cues become more distinct. Functional behavior depends on accurately anticipating impending outcomes following cues. Too little or too much

EP

stimulus generalization would decrease this accuracy, leading to maladaptive behavior.

CC

3.2 Response generalization. Just as learning about the consequences of cues requires some but not excessive stimulus

A

generalization to be functional, so too is response generalization required, particularly in instrumental situations. The precise response topography that was functional during training is often not the optimal response topography at test due to a change in response contingencies. As with stimulus generalization, the bases for this difference include changes in the environment as well as changes in the organism including the position of the organism with respect to the relevant

Maladaptive Behavior - 22 manipulandum. Responses that are not modified to fit the spatial and motoric constraints of the test situation will lead to maladapative behavior. 3.3 Generalization across other variables.

SC RI PT

Just as cues and optimal responses are subject to variation across trials due to continuing changes in the environment and the organism, so too is there ongoing variation in minor features of the outcome. Other factors that are subject to variation across trials include inter-event intervals (e.g., inter-stimulus interval and delay of reinforcement following an instrumental response) and details of the prevailing contingencies. Organisms that generalize across small

U

fluctuations in these variables but less so given larger fluctuations (in which the contingencies

N

between training and test are more apt to change) will be best served. Insufficient generalization

A

across small variations in such variables will fail to take advantage of prior learning, and excessive

M

generalization is likely to result in responding that is ineffective in a test situation that is now rather dissimilar to that of training. Both are apt to result in maladaptive behavior.

TE

D

3.4 Transfer.

The preceding discussion of generalization of learning was concerned with differences

EP

between training and test situations that were concrete surface features of stimuli and responses. Such differences are sometimes called near transfer, in contrast with far transfer which refers to

CC

generalization of rules and abstractions (i.e., deep structure) independent of specific cues and responses (e.g., Barnett & Ceci, 2002). However, we (among others) suggest that there is not a

A

sharp distinction between the two types of transfer, as transfer of abstractions and rules are often found to be positively correlated with feature similarity across tasks (Gentner & Medina, 1998; Medin & Schaffer, 1978; Wasserman & DeVolder, 1993). Like near transfer (i.e., generalization), far transfer is beneficial when the contingencies of the training task also hold for the test task, and it is maladaptive when the contingencies of the training task do not hold at test. Commonly, as the training and test situations become less similar, the contingencies of training are less apt to hold in

Maladaptive Behavior - 23 the test situation. Hence, owing to both natural selection in prior generations and prior experience of the organism in question, less far transfer of training is observed as the training and test tasks become less similar (e.g., Weson’s [1968] classic demonstration that people are poor at abstracting rules despite being good at applying them in the specific situation in which the rule was learned),

SC RI PT

which is usually, but not always, functional. Too little transfer when the same contingency holds is maladaptive and too much transfer when the same contingency does not hold is maladaptive. The tendency of an animal to exhibit far transfer, as with generalization, likely reflects natural selection of genetically encoded predispositions as well as the animal’s prior experience with related but distinctively different tasks (i.e., far transfer). Certainly, across species there are varying degrees

U

of predispositions toward far transfer across different abstractions, and training can, to at least

N

some degree, increase far transfer of the specific abstraction in question (Gentner, Loewenstein, &

A

Thompson, 2003; Wasserman, Hugart, & Kirkpatrick-Steger, 1995; Zentall et al. 2008). But to

M

date we know very little about whether training can influence any predisposition for far transfer beyond the specific abstractions on which far transfer has been trained. That is, does learning to

D

transfer one form of abstraction facilitate transfer of additional abstractions?

TE

3.5 Associative interference.

EP

As previously stated, maladaptive behavior can arise from insufficient or excessive generalization of prior concrete learning (i.e., near transfer) or from insufficient or excessive far

CC

transfer of previously learned abstractions and rules. Here ‘insufficient’ or ‘excessive’ refers to the resultant fit of prior learning to the contingencies prevailing at test. An additional factor that can

A

induce maladaptive behavior is when there is (and there often is subtly if not overtly) one or more alternative (i.e., nontarget) associations that share elements (e.g., cues, outcomes, or responses) with the target behavior assessed at test. In these cases, if the stimuli present at test activate the nontarget association(s), such activation might interfere with retrieval of the target association or elicit behavior that competes with the target response. Failures to perform in such situations

Maladaptive Behavior - 24 would take the form of insufficient generalization of the target information, not by virtue of the excessive differences between the circumstances of target training and testing, but by virtue of the differences between circumstances of nontarget training and testing being less than the differences between target training and testing (Nairne, 2002; Poirier et al., 2012).

As there are often

SC RI PT

nontarget stimulus-stimulus, stimulus-response, and response-outcome overt or subtle associations that share stimulus, response, or outcomes with target associations, associative interference frequently contributes to maladaptive behavior. Thus, many failures of generalization of learning (i.e., near transfer) can be understood in terms of associative interference (Polack, Jozefowiez, & Miller, 2017). Far transfer might (or might not) better be viewed as simply too little

U

reactivation of the target memories.

A

N

4 Instincts vs. behavioral plasticity.

M

Evolution creates predispositions for selected behaviors across generations, whereas reinforcement selects for behaviors (and predispositions towards select behaviors) within Genetically encoded predispositions reduce or eliminate the need for specific

D

generations.

TE

training. When no specific experience is needed such that the stimulus-response linkage in question appears as a function of only maturation, we say that the behavior is ‘instinctual.’ But

EP

such labeling can be misleading because the instinct vs. learning distinction is better viewed as a continuum than as two sharply distinguishable categories of behavior. When the predisposition

CC

takes advantage of the reinforcement contingencies prevailing at test, such instinctual responding spares the animal of the time and effort of learning by consequence, and as such is highly

A

functional.

But when the contingencies at test differ from those of the EEA in which the

predisposition for the instinctual behavior was selected, the instinct will lead to maladaptive behavior as discussed in Section 2.1. The benefits of ‘instinct’ over learning can be considerable when environmental contingencies are relatively stable across generations. But the same predispositions that support

Maladaptive Behavior - 25 instincts will yield maladaptive behavior by retarding or preventing learning when environmental contingencies have changed between an organism’s EEA and current circumstances. In contrast, high degrees of behavioral plasticity are apt to be beneficial with respect to reinforcement contingencies that change within the life span of an organism or change relative to those that

SC RI PT

prevailed for the organism’s recent ancestors. 5 Addressing maladaptive behaviors.

The present taxonomy of sources of maladaptive behaviors offers some implications for application.

For example, consider formal education of humans.

A major challenge in our

U

schools has been getting students to generalize what they have learned to new situations. This is

N

especially a problem when training is of abstractions and rules. That is, although we sometimes

A

observe under-generalization with respect to concrete stimulus attributes (i.e., generalization

M

decrement), we more often see failures of far transfer (Woodworth & Thorndike, 1901; Reed, Ernst, & Banerji, 1974). Not only are principles difficult to teach to people, but once they have been

D

learned, people frequently do not transfer the abstractions to very different situations. Efforts to

TE

train people to exhibit more far transfer has historically been only marginally successful. Whether that is because limited far transfer is largely genetically predisposed or because we have not yet

EP

figured out how to encourage far transfer is not clear. But recently a few researchers have claimed to have identified some effective means of encouraging both greater generalization and far transfer

CC

(e.g., Barnett & Ceci, 2002; Kurtz, Boukrina, & Gentner 2013). Among them are training with many diverse instances (Wright et al., 2016), training in diverse contexts (Gunther, Denniston, &

A

Miller, 1998; Medin & Schaffer, 1978), and practice at retrieving the target information (Karpicke & Roediger, 2008), as well as such classic manipulations as long and varied intervals between training trials (Storm, Bjork, & Bjork, 2010). Research has also identified some techniques for reducing associative interference. Associative interference is a function of (a) the relative strengths of the target and nontarget

Maladaptive Behavior - 26 memories, (b) the relative times from acquisition of the target and nontarget memories to test, and (c) the relative similarity of the contextual and punctate cues of test to those of target and nontarget training. Thus, efforts to reduce associative interference should aim to (a) train target memories to substantially greater degree than potentially interfering (i.e., similar) nontarget memories, (b) train

SC RI PT

target memories separated in time from potentially interfering memories, and (c) to the extent possible train target memories in the presence of punctate and contextual stimuli that will also be present at test while avoiding acquisition of potentially interfering memories. 6 Conclusions.

U

There are at least four central sources of seemingly maladaptive behaviors. These include

N

(a) changes in reinforcement contingencies from the EEA in which the organism’s behavioral

A

predispositions were selected, (b) changes in contingencies from the organism’s own past

M

experience, (c) behaviors that are genetically linked to beneficial traits, and (d) behaviors that were viewed as maladaptive erroneously because their beneficial consequences went unnoticed. In

D

the first two cases, maladaptive behavior arises from generalizing a training situation to a test

TE

situation in which different reinforcement contingencies prevail or from a lack of generalization between two situations in which the same reinforcement contingencies prevail. Even when the

EP

target training cues and test cues are similar, associative interference can contribute to maladaptive responding when the test cues better resemble the cues of nontarget events than the

CC

target training cues. Recognizing the roots of such maladaptive behaviors may suggest ways in

A

which their frequency can be reduced or at least compensated for.

Maladaptive Behavior - 27 References Anderson, M. C., Bjork, E. L., & Bjork, R. A. (2000). Retrieval-induced forgetting: Evidence for

a

recall-specific

mechanism. Psychonomic

Bulletin

&

Review, 7,

522-530.

DOI:

SC RI PT

10.3758/BF03214366 Anselme, P., Robinson, M. J. F., & Berridge, K. C. (2013). Reward uncertainty enhances incentive salience attribution as sign-tracking. Behavioural Brain Research, 238, 53-61. doi.org/10.1016/j.bbr.2012.10.006

Arkes, H. R., & Ayton, P. (1999). The sunk cost and Concorde effects: Are humans less than

lower

animals?.

Psychological

125,

591-600.

DOI:

A

N

10.1037/0033-2909.125.5.591

bulletin,

U

rational

Armelagos, G. J. (2014). Brain evolution, the determinants of food choice, and the

D

10.1080/10408398.2011.635817

M

omnivore's dilemma. Critical Reviews in Food Science and Nutrition, 54, 1330-1341. DOI:

EP

10.1086/276408

TE

Baldwin, J. M. (1896). A new factor in evolution. American Naturalist, 30, 441-451. DOI:

Balsam, P. D., & Silver, R. (1994). Behavioral change as a result of experience: Toward

CC

principles of learning and development. In J. A. Hogan and J. J. Bolhuis (Eds.), Causal mechanisms of behavioural development (pp. 327-357). Cambridge, UK: Cambridge University

A

Press.

Barnett, S. M., & Ceci, S. J. (2002). When and where do we apply what we learn?: A

taxonomy

for

far

transfer. Psychological

Bulletin, 128,

612-637.

DOI:

10.1037/0033-2909.128.4.612 Bertram, B. C. (1975). Social factors influencing reproduction in wild lions. Journal of

Maladaptive Behavior - 28 Zoology, 177, 463-482. DOI: 10.1111/j.1469-7998.1975.tb02246.x Bjork, R. A. (1989). Retrieval inhibition as an adaptive mechanism in human memory. In Roediger, H. L. and F. I. M. Craik (Eds.), Varieties of memory & consciousness: Essays in honour

SC RI PT

of Endel Tulving, pp. 309-330. Hillsdale, NJ: Erlbaum. Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian learning. Psychological Bulletin, 114, 80-99. DOI: 10.1037/0033-2909.114.1.80

Brown, J. S. (1969). Factors affecting self-punitive locomotor behavior. In B. A. Campbell &

U

R. M. Church (Eds.), Punishment and aversive behavior. New York: Appleton-Century-Crofts.

N

Brown, P. L., & Jenkins, H. M. (1968). Auto‐shaping of the pigeon's key‐peck. Journal of the

A

Experimental Analysis of Behavior, 11, 1-8. DOI: 10.1901/jeab.1968.11-1

from

the

influence

of

stimulation. Psychological

Review, 33,

51-58.

DOI:

D

10.1037/h0074309

external

M

Carmichael, L. (1926). The development of behavior in vertebrates experimentally removed

TE

Chow, J. J., Smith, A. P., Wilson, A. G., Zentall, T. R., & Beckmann, J. S. (2017).

EP

Suboptimal choice in rats: Incentive salience attribution promotes maladaptive decision-making.

CC

Behavioural Brain Research, 320, 244-254. doi.org/10.1016/j.bbr.2016.12.013 Cook, M., & Mineka, S. (1989). Observational condition of fear to fear-relevant versus

A

fear-irrelevant stimuli in rhesus monkeys. Journal of Abnormal Psychology, 98, 448-459. DOI: 10.1037/0021-843X.98.4.448 Darwin, C. (1871). The decent of man, and selection in relation to sex. Princeton, NJ: Princeton University Press. Davis, M. I., & Clark, D. M. (1998). Thought suppression produces a rebound effect with

Maladaptive Behavior - 29 analogue post-traumatic intrusions. Behaviour Research and Therapy, 36, 571–582. DOI: 10.1016/S0005-7967(98)00051-5 de Ghett, V. J. (1978). The ontogeny of ultrasound production in rodents. In G. M.,

aspects (pp. 343-365). New York: Garland.

SC RI PT

Burghardt and M. Bekoff (Eds.). The development of behavior: Comparative and evolutionary

Falk, J. L. (1961). Production of polydipsia in normal rats by an intermittent food schedule. Science, 133, 195-196. DOI: 10.1126/science.133.3447.195

U

Festinger, L. (1957). A theory of cognitive dissonance. Evanston, IL: Row, Peterson..

N

Foree, D. D., & LoLordo, V. M. (1973). Attention in the pigeon: Differential effects of

A

food-getting versus shock-avoidance procedures. Journal of Comparative and Physiological

M

Psychology, 85, 551-558. DOI: 10.1037/h0035300

Garcia, J., Hankins, W. G., & Rusiniak, K. W. (1974). Behavioral regulation of the milieu

D

interne in man and rat. Science, 185, 824-831. DOI: 10.1126/science.185.4154.824

analogical

encoding. Journal

of

Educational

Psychology, 95,

393-408.

DOI:

EP

for

TE

Gentner, D., Loewenstein, J., & Thompson, L. (2003). Learning and transfer: A general role

10.1037/0022-0663.95.2.393

CC

Gentner, D., & Medina, J. (1998). Similarity and the development of rules. Cognition, 65,

A

263-297. DOI: 10.1016/S0010-0277(98)00002-X Gert van Dijk, J. (2003). Fainting in animals. Clinical Autonomic Research, 13, 247–255.

DOI: 10.1007/s10286-003-0099-1 Goodwin, B. C., Browne, M., & Rockloff, M. (2015). Measuring preference for supernormal over natural rewards: a two-dimensional anticipatory pleasure scale. Evolutionary Psychology, 13, 1-11. DOI: 10.1177/1474704915613914.

Maladaptive Behavior - 30 Gunther, L. M., Denniston, J. C., & Miller, R. R. (1998). Conducting exposure treatment in multiple contexts can prevent relapse. Behaviour Research and Therapy, 36, 75-91. DOI: 10.1016/S0005-7967(97)10019-5

SC RI PT

Gwinn, G. T. (1949). The effects of punishment on acts motivated by fear. Journal of Experimental Psychology, 39, 260-269. DOI: 10.1037/h0062431

Hearst, E., & Jenkins, H. M. (1974). Sign-tracking: The stimulus-reinforcer relation and directed action. Austin, TX: Psychonomic Society.

U

Kahneman, D. (2013). Think, fast and slow. New York, NY: Farrar, Straus and Giroux.

Journal

of

the

Econometric

Society,

47,

263-291.

DOI:

A

risk. Econometrica:

N

Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under

M

10.1142/9789814417358_0006

Karpicke, J. D., & Roediger, H. L. (2008). The critical importance of retrieval for

D

learning. Science, 319, 966-968. DOI: 10.1126/science.1152408

as

amounts

increase. Psychonomic

Bulletin

&

Review, 3,

100-104.

DOI:

EP

decrease

TE

Kirby, K. N., & Maraković, N. N. (1996). Delay-discounting probabilistic rewards: Rates

10.3758/BF03210748

CC

Krebs, J. R., Kacelnik, A., & Taylor, P. (1978). Test of optimal sampling by foraging great

A

tits. Nature, 275, 27-31. DOI: 10.1038/275027a0 Kurtz, K. J., Boukrina, O., & Gentner, D. (2013). Comparison promotes learning and

transfer of relational categories. Journal of Experimental Psychology: Learning, Memory, and Cognition, 39, 1303-1310. DOI: 10.1037/a0031847 La Frenière, P. (2010). Adaptive origins: Evolution and human development. New York, NY: Psychology Press.

Maladaptive Behavior - 31 Lenoir, M., Serre, F., Cantin, L., & Ahmed, S. H. (2007). Intense sweetness surpasses cocaine reward. PloS One, 2, e698. DOI: 10.1371/journal.pone.0000698 Lieberman, L. S. (2016). Objective and subjective aspects of the drive to eat in obesogenic

across the lifespan (pp. 195-230). New York, NY: Springer.

SC RI PT

environments. In L. L. Sievert & D. E. Brown (Eds.), Biological measures of human experience

Locurto, C. M., Terrace, H. S., & Gibbon, J. (Eds.) (1981). Autoshaping and conditioning theory. New York, NY: Academic Press.

Lorenz, K. (1935). Der kumpan in der umwelt des vogels. Journal für Ornithologie, 83,

N

U

289-413. DOI: 10.1007/BF01905572

A

Luria, A. R. (1968). The mind of a mnemonist: A little book about a vast memory (L.

Medin,

D.

L.,

M

Solotaroff, Trans.). New York, NY: Basic Books. &

Schaffer,

M.

M.

(1978).

Context

theory

of

classification

D

learning. Psychological Review, 85, 207-238. DOI: 10.1037/0033-295X.85.3.207

TE

Melvin, K. B. (1971). Vicious circle behavior. In H. D. Kimmel (Ed.), Experimental

EP

psychopathology: Recent research and theory (pp. 96-115). New York: Academic Press. Miller, R. R., & Matute, H. (1998). Competition between outcomes. Psychological

CC

Science, 9, 146-149. DOI: 10.1111/1467-9280.00028

A

Nairne, J. S. (2002). The myth of the encoding-retrieval match. Memory, 10, 389-395. DOI:

10.3758/s13421-011-0133-9 Öhman, A., & Mineka, S. (2001). Fears, phobias, and preparedness: Toward an evolved module

of

fear

and

fear

10.1037/0033-295X.108.3.483

learning.

Psychological

Review,

108,

483–522.

DOI:

Maladaptive Behavior - 32 Packer, C., & Pusey, A. E. (1983). Adaptations of female lions to infanticide by incoming males. American Naturalist, 121, 716-728. DOI: 10.1086/284097 Patitucci, E., Nelson, A. J. D., Dwyer, & Honey, R. C. (2016). The origins of individual

Experimental

Psychology:

Animal

Learning

SC RI PT

differences in how learning is expressed in rats: A general-process perspective. Journal of and

doi.org/10.1037/xan0000116

Cognition,

42,

313-324.

Pawlowski, B., Atwal, R., & Dunbar, R. I. M. (2008). Sex differences in everyday risk-taking behavior in humans. Evolutionary Psychology, 6, 29-42. DOI: 10.1177/147470490800600104

U

Poirier, M., Nairne, J. S., Morin, C., Zimmermann, F. G., Koutmeridou, K., & Fowler, J.

N

(2012). Memory as discrimination: A challenge to the encoding–retrieval match principle. Journal of

A

Experimental Psychology: Learning, Memory, and Cognition, 38, 16-29. DOI: 10.1037/a0024956

M

Polack, C. W., Jozefowiez, J., & Miller, R. R. (2017). Stepping back from ‘persistence and

TE

10.1016/j.beproc.2017.03.014

D

relapse’ to see the forest: Associative interference. Behavioural Processes, 141, 128-136. DOI:

Polack, C. W., McConnell, B. L., & Miller, R. R. (2013). Associative foundation of causal

EP

learning in rats. Learning & Behavior, 41, 25-41. DOI: 10.3758/s13420-012-0075-5

CC

Reed, S. K., Ernst, G. W., & Banerji, R. (1974). The role of analogy in transfer between

A

similar problem states. Cognitive Psychology, 6, 436-450. DOI: 10.1016/0010-0285(74)90020-6 Reiss, S. (1980). Pavlovian conditioning and human fear: An expectancy model. Behavior

Therapy, 11, 380-396. DOI: 10.1016/S0005-7894(80)80054-2 Seligman, M. E. P. (1971). Phobias and preparedness. Behavioral Therapy, 2, 307-20. DOI: 10.1016/S0005-7894(71)80064-3 Seligman, M. E. P. & Hager, J. L. (Eds) (1972). Biological boundaries of learning. New York,

Maladaptive Behavior - 33 NY: Appleton-Century-Crofts. Serjeant, G. R. (2010). One hundred years of sickle cell disease. British Journal of Haematology, 151, 425-429. DOI: 10.1111/j.1365-2141.2010.08419.x

SC RI PT

Sidman, M. (1953). Two temporal parameters of the maintenance of avoidance behavior by the white rat. Journal of Comparative and Physiological Psychology, 46, 253-261. DOI: 10.1037/h0060730

Sidman, M., & Fletcher, F. G. (1968). A demonstration of auto‐shaping with monkeys. Journal

of

the

Experimental

Analysis

Behavior, 11,

307-309.

DOI:

N

U

10.1901/jeab.1968.11-307

of

A

Silva, F. J., Silva, K. M., & Pear, J. J. (1992). Sign- versus goal-tracking: Effects of

M

conditioned-stimulus-to-unconditioned-stimulus distance. Journal of the Experimental Analysis of Behavior, 57, 17-31. doi.org/10.1901/jeab.1992.57-17

D

Sivak, J (2012). The cause(s) of myopia and the efforts that have been made to prevent it.

TE

Clinical and Experimental Optometry. 95, 572–82. doi:10.1111/j.1444-0938.2012.00781.x.

EP

Soares, J. S., Polack, C. W., & Miller, R. R. (2016). Retrieval-induced versus context-induced forgetting: Does retrieval-induced forgetting depend on context shifts? Journal of Psychology:

Learning,

Memory,

and

Cognition, 42,

366-378.

DOI:

CC

Experimental

A

10.1037/xlm0000171 Storm, B. C., Bjork, R. A., & Storm, J. C. (2010). Optimizing retrieval as a learning event:

When and why expanding retrieval practice enhances long-term retention. Memory & Cognition, 38, 244-253. DOI: 10.3758/MC.38.2.244 Swithers, S. E., Davidson, T. L. (2008). A role for sweet taste: Calorie predictive relations in energy

regulation

by

rats.

Behavioral

Neuroscience,

122,

161-173.

DOI:

Maladaptive Behavior - 34 10.1037/0735-7044.122.1.161 Tinbergen, N., & Perdeck, A. C. (1951). On the stimulus situation releasing the begging response in the newly hatched herring gull chick (Larus argentatus argentatus Pont.). Behaviour, 3,

SC RI PT

1-39. DOI: 10.1163/156853951x00197 Wasserman, E. A., & DeVolder, C. L. (1993). Similarity-and nonsimilarity-based conceptualization

in

children

and

pigeons. Psychological

10.1007/BF03395912

Record, 43,

779-793.

DOI:

Wasserman, E. A., Hugart, J. A., & Kirkpatrick-Steger, K. (1995). Pigeons show

U

same-different conceptualization after training with complex visual stimuli. Journal of Experimental

A

N

Psychology: Animal Behavior Processes, 21, 248-252. DOI: 10.1037/0097-7403.21.3.248 Watson, J.B., Rayner, R. (1920). Conditioned emotional reactions. Journal of Experimental

M

Psychology, 3, 1–14. DOI:10.1037/h0069608

D

Williams, D. R., & Williams, H. (1969). Auto‐maintenance in the pigeon: Sustained pecking

TE

despite contingent non-reinforcement. Journal of the Experimental Analysis of Behavior, 12,

EP

511-520. DOI: 10.1901/jeab.1969.12-511 Woods, S. C., & Begg, D. P. (2015). Regulation of the motivation to eat. In E. H. Simpson &

CC

P. D. Balsam (Eds.), Behavioral neuroscience of motivation (pp. 15-34). New York, NY: Springer. Woodworth, R. S., & Thorndike, E. L. (1901). The influence of improvement in one mental

A

function upon the efficiency of other functions. Psychological Review, 8, 247-261. DOI: 10.1037/h0074898 Wright, A. A., Magnotti, J. F., Katz, J. S., Leonard, K., & Kelly, D. M. (2016). Concept learning set‐size functions for Clark's nutcrackers. Journal of the Experimental Analysis of Behavior, 105, 76-84. DOI: 10.1002/jeab.174

Maladaptive Behavior - 35 Xu, F., Spelke, E. S., & Goddard, S. (2005). Number sense in human infants. Developmental Science, 8, 88-101. DOI: 10.1111/j.1467-7687.2005.00395.x Zentall, T. R. (2016). Resolving the paradox of suboptimal choice. Journal of Experimental

SC RI PT

Psychology: Animal Learning and Cognition, 42, 1-14. DOI: 10.1037/xan0000085 Zentall, T. R., Wasserman, E. A., Lazareva, O. F., Thompson, R. K. R., & Rattermann, M. J. (2008). Concept learning in animals. Comparative Cognition & Behavior Reviews, 3, 13-45. DOI:

A

CC

EP

TE

D

M

A

N

U

10.1002/jeab.55