KINDS OF BEHAVIOUR
Robert Aunger* and Valerie Curtis*
* The Hygiene Centre
London School of Hygiene and Tropical Medicine
London WC1E 7HT
email@example.com (corresponding author), firstname.lastname@example.org.
‘The confusion and barrenness of psychology cannot be explained by
calling it a ‘young science’…in psychology there are experimental methods and
conceptual confusion.’ Wittgenstein, Philosophical Investigations.
IN PRESS, BIOLOGY AND PHILOSOPHY
Sciences able to identify appropriate analytical units for their domain, their natural kinds, have
tended to be more progressive. In the biological sciences, evolutionary natural kinds are
adaptations that can be identified by their common history of selection for some function. Human
brains are the product of an evolutionary history of selection for component systems which
produced behaviours that gave adaptive advantage to their hosts. These structures, behaviour
production systems, are the natural kinds that psychology seeks. We argue these can be identified
deductively by classing behaviour first according to its level of behavioural control. Early animals in
our lineage used only reactive production, Vertebrates evolved motivation, and later Primates
developed executive control. Behaviour can also be classified by the type of evolutionary benefit it
bestows: it can deliver either immediate benefits (food, gametes), improvements in the individual’s
position with respect to the world (resource access, social status), or improvements in the ability to
secure future benefits (knowledge, skill). Combining history and function implies the existence of
seven types of behaviour production systems in human brains responsible for reflexive, instinctual,
exploratory, driven, emotional, playful and planned behaviour. Discovering scientifically valid
categories of behaviour can provide a fundamental taxonomy and common language for
understanding, predicting and changing behaviour, and a way of discovering the organs in the
brain – its natural kinds – that are responsible for behaviour.
Key words: evolution; behaviour; emotion; brain; cognition; natural kinds
Fruitful scientific investigation generally requires that entities, phenomena or processes be placed
in meaningful classes called ‘natural kinds’. Without such categorisation, empirical generalization,
explanation and prediction can fail. (Allen & Bekoff, 1997; Dupré, 2000; Griffiths, 1997; LaPorte,
2004; Quine, 1969) Modern chemistry, for example, derives its explanatory power from the
recognition that substances are composed of a small number of different, but unchangeable kinds
of atoms. Its predecessor, alchemy, failed as a science because it assumed that a mysterious
substance, the philosopher’s stone, could transmute metals. As a result, alchemy could not
correctly identify elements, or predict their immutability. Similarly, evolutionary biology could not
explain the inheritance of information until genes, the correct units of inheritance, were described.
(Wagner & Wagner, 2003) Atoms and genes are basic building blocks of their respective sciences,
their natural kinds. (Wagner, 1996) Instances of a natural kind can be meaningfully grouped
together because they share something fundamentally real, which can be discovered through
objective investigation. (Boyd, 1991; Millikan, 1984) Science based on natural kinds thus allows us
to map what matters in the world, giving us understanding and, hence, power over it. (Sterelny,
Psychology is the science of mental processes and behaviour. (Editors, 2006; Myers, 2006)
However, psychology and the behavioural sciences have still not identified or agreed upon a set of
component natural kinds. Scholars have argued that the brain is designed to perform functions
such as: computation (Mountcastle, 1998), learning (Dayan & Abbott, 2001), the detection of
discrepancies, (Barlow, 1994) reasoning (Johnson-Laird, 2006) or predicting the future (Hawkins &
Blakeslee, 2004). However, from an evolutionary perspective, the primary function of the brain is to
produce adaptive behaviour. (Churchland & Sejnowski, 1992; Freeman, 1999; Hebb, 1949; Newell,
1990; Skinner, 1938) Attention, learning, reasoning and prediction are of no use unless they
facilitate the production of behaviours that help an animal stay alive a bit longer or reproduce itself.
Animals are the only major kingdom of life to have adopted behaviour as their primary means of
adaptation (Lorenz, 1965; Tinbergen, 1951) and they alone have brains. Further, those animals
that live in more complex environments tend to have more complex brains to deal with them.
(Godfrey-Smith, 1996) Since brains serve to produce behaviour, the natural kinds of psychology
and the behavioural sciences must be the units in the brain that produce behaviour.
In biology, natural kinds such as cells or organs can be distinguished by their similar features.
These features are shared because they result from a common history of natural selection for
serving a particular purpose. Thus animal brains are a natural kind because they share a common
phylogenetic history and a common function, that of producing behaviour. (Haslam, 2002)
Similarly, neurons can be considered a natural kind because we know that they evolved to transfer
information within brains. However, behaviour is controlled by structures which lie between the
levels of individual neurons and of whole brains.
How, then, can we identify the component organs of the brain? To do so we have to be able to
both identify its regular structures and describe their adaptive functions. Brains can only be
meaningfully carved up by referring to the process that shaped behaviour, and hence brains, in the
first place -- that of evolution. Armed with such a classification, neuroscientists should be able to
identify the structures responsible for producing these categories of behaviour within brains. These
behaviour production units would then be the natural kinds that we seek.
We begin this paper by setting out what is meant by a natural kind in biology, and show how
evolutionary kinds arise through a shared history of adaptation for function. We then set out a
classification of animal behaviour by adaptive function. We propose that there are only three kinds
of evolutionary function for behaviours: those that improve an organism’s physiological state, those
that improve its state with respect to the world, and those that allow it to improve its abilities to
acquire evolutionary benefits. Evolution has invented three ways of carrying out those functions –
in historical sequence, through reactive, motivated and executive behaviours. When combined,
these conditions provide us with seven distinct classes of behaviour which we label reflex, instinct,
exploration, drive, emotion, play, and planning. We have italicised our usage of these words to
emphasise that we have assigned them principled new meanings, different from previous, often
contested, histories of lay and scientific use. The definitions of these coinings, along with those of
related terms, can be found in Table 1. We predict that these seven classes of behaviour will be
produced by corresponding mechanisms, or behaviour production units, in the brain, which are its
natural kinds for behaviour. In the discussion we reflect on the pitfalls of the approach we have
taken and look at its practical implications both for neuroscience and for studying, predicting and
Table 1 Glossary of terms
CONCEPT DEFINITION EXAMPLES
Behaviour Self-propelled movement producing a
functional interaction between an animal
and its environment
Build shelter, Avoid
predation, Hunt with
An evolved psychological mechanism for
producing an optimized response to a cue,
situation or goal
A resource tightly correlated with
increases in biological fitness Food,
Cue A signal of some evolutionarily salient
state or variable (environmental or
Low blood glucose
Action Behavioural activity produced in response
to a cue resulting from reactive-level
Reflex A class of BPUs and behaviours triggered
by a cue and designed to attain a
Instinct A class of BPUs and behaviours triggered
by a cue and designed to attain a
Exploration Behaviour produced by a class of BPUs
triggered by default and designed to attain
an aptitudinal end-state through reactive
Reward The value of feedback from behaviour
(may be expressed as deviation from
Indicator A mental representation that indicates the
degree of discrepancy between the current
state of some variable (based on a cue)
and its optimal state
Motivation A psychological state that arises when an
indicator is greater than some threshold
Need A task related to an evolutionarily
significant aspect of an animal’s ecological
niche which requires goal-directed
behaviour to solve
Invest in pair-bond,
Maximize social status
Episode A set of actions produced as a motivated
response to a cue or situation Display threat face
Drive A class of BPUs and behaviours triggered
by an indicator and designed to achieve a
Emotion A class of BPUs and behaviours triggered
by an indicator and designed to achieve a
Compete for status
Play A class of BPUs and behaviours triggered
by default and designed to achieve an
aptitudinal end-state through motivated
Goal A mental representation of an end-state Satiety,
Copulate with mate
Objective An arbitrarily distant or abstract (i.e., non-
evolved) goal Have a job
Planning A BPU designed to achieve objectives
through executive-level control ----
Plan An action sequence produced by the
planning BPU Search want ads fill
out application mail
Defining kinds of behaviour
The original conception of natural kinds, due to Aristotle and pursued by the English philosophers
Locke and Mill, viewed them as sets of things characterized by a necessary and sufficient suite of
characteristics. These characteristics are shared by all members of the kind because they are
subject to universally applicable laws. For example, water, gold and stars have the same
composition, and always do the same kinds of things, wherever and whenever they are. This
commonality is due to a shared essence which guarantees the identity of the natural kind and
serves as its principal defining element, regardless of what those instances actually look like.
(Kripke, 1972; Putnam, 1975)
This Aristotelian concept is now seen as being applicable to physical and chemical kinds, but not to
biological or social kinds, which are more restricted in scope, both physically and temporally.
(Boyd, 1991; Boyd, 1999b; Griffiths, 1999; Millikan, 1999) For example, any example of water will
have exactly the same chemical constituents and structure (H
0), but all members of the human
species do not have exactly the same composition and structure (e.g., some are male, others
female). Conversely, there may be an organism on the planet Venus with exactly the same
composition and structure of a human, which would not qualify as human because it does not have
the same history (Ghiselin, 1974; Hull, 1978). The objects of theory in the biological and social
sciences are thus historical in ways that the objects of physical and chemical sciences are not. As
a result, biological and social kinds are ‘fuzzier’, and are subject to qualified laws. (Hacking, 2002;
Millikan, 1999; Millikan, in press; Rieppel, 2005b)
The homeostatic property cluster (HPC) concept of natural kinds (Boyd, 1991; Boyd, 1999a) is
now widely taken to be the best foundation for biological and social kinds. (Charland, 2002;
Griffiths, 1999; Keller, Boyd & Wheeler, 2003; Kornblith, 2002; Millikan, 1999; Rieppel, 2005b) HPC
kinds are ‘homeostatic’ because some force causes deviations in the qualities of members to
return toward a central tendency. (Boyd, 1991; Boyd, 1999b) HPC kinds are the result of the same
causal force acting on all of its members. While the Aristotelian, essentialist position assumes that
the members of a kind are similar due to a shared essence or intrinsic property, HPC theory allows
a broad range of properties and mechanisms jointly to constitute the kind. (Mallon, 2003) For
example, species are subject to natural selection pressures from environmental factors which
cause members of that species to have similar qualities (e.g., morphological regularities). A
member of an HPC kind does not have to share every feature of its kind with all of its brethren;
instead, individual members need only have enough of the important properties in the defining
cluster to qualify as belonging. (Boyd, 1991; Boyd, 1999a) This weaker concept of membership
seems appropriate for objects which do not share an essence, as in Aristotelian natural kinds, but
rather similar causes.
HPC kinds for which the homeostatic mechanism is common descent constitute a class which we
will call ‘evolutionary kinds’. (Charland, 2002; Griffiths, 1997; Griffiths, 2004a; Rieppel, 2005a) In
this case, the reason that members of the kind have a number of features in common is their
shared history of subjection to natural selection pressures. Evolutionary kinds thus have law-like
properties, but are limited in the range to which the kind extends, due to their historical specificity.
(Boyd, 1991; Griffiths, 2004a; Rieppel, 2005b)
The members of evolutionary kinds produce effects which biologists recognize as phenotypes.
These effects have been selected for the function that they performed in the past; those with higher
fitness value tended to be retained, while those with lower value tended to disappear. Eyes are a
quintessential example. (Fernald, 1997; Gehring & Ikeo, 1999) All eyes share a common function --
that of converting photons to signals that can be registered by nervous tissue. However, some
structures perform this function better than others. The current types of eyes within a given lineage,
such as Mammals, take the form they do as a consequence of a past history of selection for their
function and are thus an evolutionary natural kind.
Natural kinds tend to appear in nested hierarchies: in chemistry, substances are made up of
molecules, which are made up of atoms; similarly, in biology, organisms contain organs, which
contain tissues which are made of cells. Thus functional units exist at multiple levels of
organisation. (Bolker, 2000; Dawkins, 1976; Hartwell et al., 1999; Raff, 1996; Wagner, 1996) The
following section proposes how a nested hierarchy of the brain-based units responsible for
behaviour can be uncovered.
Evolutionary kinds and behaviour
Behaviour can be defined as self-propelled movement producing a functional interaction between
an animal and its environment, such as finding a mate or fleeing from a predator. (Millikan, 2000)
Some organisms which lived in rapidly changing environments found it to their advantage to
develop a mechanism which allowed their behavioural responses to be contingent on changing
conditions. (Allman, 1999; Godfrey-Smith, 1996; Sterelny, 2003) Brains have been selected to
produce behaviour adaptive to an animal’s circumstances because mechanisms ensuring the
production of such responses would have been favoured by selection. The result has been
regularized structures within animals that consistently produce adaptive behaviour. We call these
‘behaviour production units’ (BPUs). We argue that these BPUs constitute the fundamental natural
kind associated with behaviour.
Figure 1: Behaviour Production Model
Inputs Reactive Selection
Models of brain function in psychology and in artificial intelligence often divide the process of
behaviour production into three parts: sensory systems at the ‘front end’ for recognizing stimuli,
motor systems at the ‘back end’ for generating behaviour, and something in between, often called
‘cognition’, ‘information processing’, or ‘central processing’ (Harnish, 2001; Mesulam, 1998;
Sloman, 2001). We argue that this middle box is at least partly composed of BPUs. In higher
animals, BPUs are composed of sets of internal structures working in functional collaboration to
produce behaviour. (Gershenson, 2001; Michaud, 2002) In higher organisms like humans,
behaviour production encompasses the development of behavioural proposals by BPUs, the
selection of one proposal by some mechanism, and its execution by the peripheral nervous
system. We consider perceptions to be inputs to BPUs, while motor commands are their output.
Figure 1 presents a schematic picture of our behaviour production model. (We will later describe
the three levels of control at which BPUs can function, represented as Reactive, Motivated and
As with any evolutionary kind, these regular structures are produced, and can be distinguished, by
their history and function. That is, members of a natural kind related to behaviour must be
functionally specific, and second, they must be restricted to a class of organisms which have a
shared evolutionary history (what biological systematists call a ‘clade’, due to being on the same
branch of a phylogenetic tree (Hennig, 1966)). In what follows, we first outline the history of the
development in the means through which behaviour has been produced, and then the classes of
function which behaviour can accomplish, given these means. We shall see that this offers us a
principled ways of carving behaviour at its joints, and hence of predicting the existence of
corresponding BPUs in brains. Finally, we will observe that the types of behaviour so defined,
perhaps unsurprisingly, approximately correspond to some established, if contested, psychological
Types of behaviour production
We first must restrict our examination of behavioural kinds to the history of evolutionary
developments in a particular lineage. Psychologists are primarily concerned with explaining
behaviour in our own species, which means we will restrict our attention to the animals of which we
are the direct descendants.
As with any historical phenomenon, the way in which things work can change over time. Just as
cilia can become fins, which can evolve into legs, (Coates, Jeffrey & Ruta, 2002) so early
mechanisms in brains for producing beneficial behaviour can be elaborated and improved upon in
later descendants. Brains have shown remarkable increases in size and complexity in the lineage
between early multi-cellular animals and our group of Primates. (Allman, 1999; Butler & Hodos,
2005; Northcutt, 2002; Streidter, 2005) What could have been the selective pressure responsible
for this increase in investment in brains by later species in our lineage? With respect to behaviour,
we believe that the advantage of devoting greater neural resources to the production of behaviour
is that it increases the time horizon over which responses can be controlled. We therefore suggest
that there have been selective pressures on some lineages to increase the time span over which
behavioural response is controlled, ranging from an instant reaction to complex and carefully
planned sequences of behaviour. If an animal’s time horizon is short (due to limitations of memory
or information processing), then its behaviour at time t is purely responsive to stimuli at time t-1; if
its horizon is longer, then it can ignore a stimulus at time t-1 to meet a need formed at time t-2, thus
performing a different behaviour at time t than if it had been purely responsive. With long-term
behavioural control, an organism can commit itself at an arbitrary time (i.e., time t-n) to ignore
stimuli arising during the intervening period in order to achieve a desirable end-state at time t.
Animals which can forgo a present benefit can thus enjoy a higher return over the longer term by
pursuing a different sequence of behaviours than would be dictated by responding to current
conditions at each moment. The ability to calculate responses over longer time-frames can thus
increase the adaptive value of behaviour.
It is also the case that the architecture of the human brain reflects its evolutionary history -- some
control mechanisms are similar to those present in unicellular animals, some structures are similar
to those of Invertebrates, and others are Mammalian in origin (Butler & Hodos, 2005; Rubenstein et
al., 1999; Streidter, 1998).This gives us a reasonable expectation that there have been historical
developments in the ways that animal brains work. The means for producing behaviour should
therefore be cobbled together by evolution out of pre-existing parts, so that descendant species in
a lineage tend to have the qualities of their ancestors, with a few changes. We thus argue that
there have been three points at which the degree of control over behaviour production made
significant advances in the lineage leading to humans (Table 2).
Table 2: Types of Behaviour Production Units
Simple Reactive Metazoans
(hydra) cues ---- short
(bony fish) goals representation
Complex Executive Primates
* Includes time before response (from time of stimulus), future time horizon over which response is
calculated prior to enactment, and length of time over which behavioural response is controlled
(see behavioural phenotype)
The simplest units of behaviour production are probably composed of one or more neural circuits.
These units can be associated with parts of the simple nerve nets of the earliest multicellular
animals, the Metazoans. The behaviour produced by these BPUs can be described as reactive,
because cues to evolutionarily significant situations are recognised and appropriate actions are
engaged almost immediately (i.e., in response to conditions at time t-1).
As brains become more complex, they show evidence of increasingly hierarchical structure. (Cziko,
1995; Geary & Huffman, 2002; Quartz, 2001; Swanson, 2003) The first animals to have a
hierarchical organisation to their nervous systems were the Vertebrates, which have brains with
multiple centralized ganglia as well as peripheral nerve circuits. (Streidter, 2005; Swanson, 2003)
We postulate that units of control at this stage are composed of BPUs with intermediate degrees of
complexity. These could divorce behaviour from immediate responses to cues, and hence get
greater benefits (i.e., by considering conditions at t-2). In effect, animals with this level of control
can persist in a particular behaviour because they are pursuing a goal, or particular end-state.
(Austin & Vancouver, 1996) This motivation provides a means of prioritisation, permitting an animal
to keep to a course of action, in the face of cues to respond otherwise. (Berridge, 2004; Deci, 1975;
Franken, 2001; Wong, 2000)
Motivated animals are responding to what we call an ‘indicator’. Indicators are highly constrained,
often subconscious mental representations of an abstract state which are based on inputs from the
body and the environment. For example, hunger is a psychological state (which can be felt in
humans) triggered by the level of grehlin in the blood, (Druce, Small & Bloom, 2005) and by the
presence of food and regulated by levels of gastric distension. (Gibbs, Maddison & Rolls, 1981)
Indicators evolved to provide animals with information about the fact that their current situation
deviates in some evolutionarily significant way from what is selectively advantageous. What
indicators evolved to signal is a state of need – in particular, the need to solve an ecological
problem. (Sterelny, 2003) Needs are related to some dimension of an animal’s niche, such as its
feeding or reproductive strategy, which presents an animal with a problem it must solve in order to
survive or reproduce.
Motivated behaviour is typically manifested in episodes of goal-directed behaviour that lead to the
extinction of the indicator through the meeting of a need. (Berridge, 2004; Franken, 2001; Murray,
1938) Indicators are constantly being calculated by the brain’s attention or vigilance system, based
on cues coming in from the environment or body. If an indicator exceeds its threshold value,
motivation occurs, activating the relevant BPU to calculate behavioural options that will reduce the
indicator below its threshold value at least cost and with highest reliability. An animal then engages
in behaviour via a mechanism that selects the favoured behavioural option among the outputs of
various BPUs. Stimuli representing changes in the environment and body due to the behaviour
feed into the reward system, which calculates the value of that behaviour as a function of how
much the indicator was reduced. Reduced motivation is positively rewarded, and motivated
learning associates the behaviour with that reduction, increasing the likelihood of engaging in that
behaviour the next time the indicator exceeds its threshold.
With Primates, we argue that a third stage of development in the complexity of BPUs was
achieved. This ‘executive’ control is associated with a new kind of goal: ‘objectives’. Objectives are
mentally represented end-states which differ from indicators in being multi-dimensionally valued,
highly contextualized, and long-term. They are robust mental representations, not the simple
representation of internal (e.g., hormone) or environmental cues. Objectives probably arose when -
other individuals had to be dealt with as agents, so that mental models of conspecifics within
models of behavioural options were required. The adaptive advantage of this form of higher control
is that an indicator such as a hunger signal could temporarily be ignored -- for example, in the
pursuit of trading for a tool that would lead to a better harvest later in the year. The type of
behavioural control available to an animal with the ability to pursue objectives we call ‘executive’.
This allowed the planning of future action sequences to attain objectives through use of a new kind
of cognitive control resident in the neocortex. (Koechlin, Ody & Kouneiher, 2003; Miller & Cohen,
2001; Wood & Grafman, 2003) Executive level control can also use motivation-level mechanisms
for the hedonic valuation of imagined future states. (Gray, Braver & Raichle, 2002; Ochsner &
Gross, 2005; Roseman, 1984)
The ability to disregard otherwise relevant concerns was made possible by increasingly
sophisticated representations of the animal’s situation in the world. The earliest animals interpreted
in-coming stimuli in terms of what we narrowly define as cues. Thus reactive animals act directly on
cues with no need for internal representations – that is, no ability to re-present sensations in an
internal form for further processing.
The ability to ignore nuisance stimuli in favour of particular
indicators requires the ability to have mental representations of (some aspects of) the world. (Wood
& Grafman, 2003) Thus, motivated animals were capable of first-order mental representations, the
most important of which were need-states, body states, categories of objects in the world, and
To achieve strategic goals, Mammals evolved the ability to represent representations – that is, to
‘meta-represent’. (Hughlings Jackson, 1958; Perner, 1991; Sterelny, 1998) This ability is
associated with the neocortex, which first appeared in Mammals. (Allman, 1999; Northcutt & Kaas,
1995) The essence of meta-representation is the ability to hold a hierarchical structure of
representations in memory and manipulate it while continuing to manage the relationships between
representations, and hence maintain a consistent body of knowledge about the world. (Sperber,
2000) BPUs which involve the neocortex, and possibly other parts of the brain as well, are the most
complex which have yet evolved. Animals with complex BPUs can represent their own
representations to themselves. Some of them can even work towards objectives, rather than
reacting to cues or indicators. (Rolls, 1999) Some animals (possibly only humans) can also
objectivise their own mental states – that is, see themselves as having thoughts about their own
thoughts. (Dennett, 1996; Proust, in press; Rosenthal, 2000) This allows animals with these special
meta-representation abilities to engage in mental simulations; they have the ability to mark some
mental representations as decoupled from states of the world, which allows temporary models of
hypothetical situations to be created. (Cosmides & Tooby, 2000; Dienes & Perner, 1999; Sperber,
2000; Stanovich, 2004)
Reactive BPUs employ direct responses; unlike motivated behaviour, they do not refer to set-points
(i.e., involve indicators) or require calibration against indicators via rewards. Motivational BPUs are
iterative, constantly engaged in loops between the animal’s current state and the previous state of
the indicator. Motivated behaviour selection is based on how behaviours have been rewarded in
similar previous situations. Cognition, on the other hand, is based on simulation of future
outcomes. Unlike the iterative try-test-retry-exit system for goal achievement of motivational BPUs,
planning can rely on a calculated sequence of options to achieve valuable objectives.
These different units of behaviour production have a number of features as a consequence of their
different degrees of complexity. First, we suggest that this trend in increasing control has been
accompanied by qualitative differences in the behavioural phenotypes produced by these types of
BPUs. At first, reactions consisted of individual actions (which may be composed of multiple
events), while motivations produce relatively short sequences of actions (which we will call
episodes), composed of a number of habitual and automatic actions, often sequence-dependent,
but controlled as a unit (such as getting dressed) (Barker & Wright, 1954; Schank & Abelson,
1977). Executive control produces potentially indefinite chains of actions that may last over a
lifetime in the form of executed plans, which often accomplish multiple goals towards the
achievement of an overall objective. (Koechlin et al., 2003; Zacks & Tversky, 2001)
It seems, therefore, that there has been a historical sequence in development of the mechanisms
for the control of behaviour. Simple animals execute simple reactive behaviours in response to
cues, Vertebrates evolved the ability to ignore immediate cues and pursue goals through motivated
behaviour, and finally Primates added a new layer of executive control of behaviour which would
allow them to achieve strategic end-states and objectives through planned sequences of
These reactive, motivated and planned types of behavioural control coincide with brain structures
which have long been identified by neuroscientists (as simple circuits or columns (Hebb, 1949;
LeDoux, 2000; Mountcastle, 1957), modules (Damasio, 2003; Freeman, 2000; Panksepp, 1998;
Tooby & Cosmides, 1992) and systems (Kelso, 1995; Thelen & Smith, 1994), respectively). All
three of these structures are present in human brains. (Damasio, 2003; Rolls, 1999; Rolls, 2005)
Similar levels of control have also been widely recognized in the literature on cognitive evolution.
(Dennett, 1996; Ortony, Norman & Revelle, 2004; Rolls, 1999; Rolls, 2005; Sloman, 2001)
Functions of behaviour
Evolutionary kinds are also distinguished by their function. (Boyd, 1991; Millikan, 1989) Behaviour
serves to put animals into states which have different types of benefits in evolutionary terms. Three
categories of state can be distinguished:
• ‘physiological’ end-states, which provide changes to evolutionary benefits themselves
• ‘situational’ end-states, which produce a changed relationship with the world, such as
access to territory or status, which tend to lead to future evolutionary benefits and
• ‘aptitudinal’ end-states, which produce a changed capacity to gain future situational or
physiological benefits through the acquisition of knowledge or skill
The first category of ‘physiological’ end-states provides reproductive or survival benefits directly.
These end-states tend to be focused on the condition of the body, because gaining immediate
evolutionary benefits must involve the acquisition of some resource, or ameliorate the survival
chances of the body. It is thus behaviour that is aimed at getting resources into the body, and
wastes out (eating, drinking, excreting), keeping the body within a range of conditions for optimal
functioning (move to suitable air, temperature, light, humidity conditions), at avoiding physiological
damage (projectiles, cliff edges, parasites) or exchanging gametes (copulation).
The second way to use behaviour to get adaptive benefits is to manipulate one’s niche so as to put
oneself into a situation where the acquisition of benefits becomes more likely in future. Effort can
be directed at improving the physical world (for example, by finding or building safe, productive
habitats, or by hoarding resources), the biological world (for example by caching food or by
avoiding pathogen habitats), or the social world (for example, by investing in offspring, by investing
in a mate so they will help rear children, or by investing in improving social status as to get better
access to resources). Biologists refer to this type of behaviour as ‘niche construction’. (Odling-
Smee, Laland & Feldman, 2003)
Our third category, ‘aptitudinal’ behaviours, is even more indirectly related to evolutionary benefits.
In this case behaviour serves to improve the actor’s own skills and abilities to carry out
physiological or environmental tasks more effectively in future (for example, through practicing
skills). The focus of aptitudinal behaviour is thus on changing the state of the brain, where memory
and skill-based knowledge resides. (Deci & Ryan, 1985b; Maslow, 1943; White, 1959)
This analysis suggests, then, that adaptive behaviour can be directed at one of three kinds of end-
states; those that improve: the state of the body, the state of the world, or the state of the
behavioural control system (i.e., brain) (Table 3). These three sorts of end-state define the ways in
which behaviour can be functionally distinguished.
Our distinction between these three kinds of end-states is supported by evidence from psychology
and neuroscience. The distinction between physiological and other end-states can be seen in brain
imaging studies in humans that have shown that separate systems evaluate choices between
immediate biological rewards and those delayed in time, (McClure et al., 2004) even when the
delay is only a few minutes. (McClure et al., 2007) Different brain systems seem to be evolved in
valuing primary versus secondary rewards; there is a clear demarcation in the brain’s calculations
between immediate benefits and future benefits at any remove from the present. The result of this
bias, in behavioural terms, can be seen in paradoxes of inter-temporal choice: Faced with a choice
between consuming $10 today and $11 tomorrow, some people will choose to consume the lesser
quantity today. However, when these same individuals are faced with the choice between the same
$10 a year from now and $11 a year and a day from now (the same time difference), they choose
to wait and consume the greater quantity. (Frederich, Lowenstein & O’Donoghue, 2002) In effect,
behaviours which immediately produce psychological rewards are hyper-stimulating.
Similarly, the division between aptitudinal and other kinds of end-states is supported by the
literature on intrinsic motivation. (Deci, 1975; Ryan & Deci, 2000) Intrinsic motivation occurs when
behaviour is performed for its own sake rather than to obtain material or social reinforcement (i.e.,
extrinsic motivation). Internally motivated behaviours tend to be felt as pleasurable, and are
typically called playful, creative, or curious. (Lepper, Greene & Nisbett, 1973) Tangible rewards
tend to have a substantial negative effect on intrinsic motivation – that is, providing people with
money, food or other inducements to do a task which they find pleasure will tend to reduce their
willingness to perform it, (Deci, 1971; Deci & Ryan, 1985a; Lepper & Henderlong, 2000) or to do so
less creatively. (Amabile, 1996) This suggests that some behaviour is performed for the functional
reward of skill-based learning for learning’s sake – just the function we have postulated to
characterize aptitudinal behaviour.
Table 3: Categories of behavioural end-state
END-STATE DESCRIPTION FOCUS
Physiological Provides benefits directly (e.g., offspring, or
resources for survival) Body
Situational Changes relationship of agent to the world such
that ability to secure future evolutionary benefits
is increased World
Aptitudinal Changes capability or capacity to carry out tasks
more effectively in future Brain
Classes of behavioural kind
We have now identified two ways of classifying behaviour – first, by the degree of sophistication in
the way behaviour is controlled by brains (reactive, motivated and executive), and second, by the
end-states which it is designed to produce (physiological, situational and aptitudinal). These criteria
correspond to the two qualities which define an evolutionary kind: shared history and function. If we
put the dimension of control together with the functional differences we have identified – that is,
treat the levels of control as criteria orthogonal to the sorts of end-states achieved – the resulting
two-dimensional matrix defines what we will call ‘classes’ of behaviour. Associated with these
classes of behaviour, we argue, will be classes of BPU in the brain to produce them.
We present these classes as cells in a two-dimensional matrix in Table 4. Each class of behaviour
can be defined succinctly by its function and level of control. We suggest a set of terms for these
classes. Notice that the names we have chosen for these classes are familiar ones, such as reflex
and emotion. We have opted for this strategy because behaviours that are commonly thought of as
reflexive or emotional, for example, do mostly cluster within our new classification.
Because they have specific, succinct definitions, each class of behaviour can be easily described
(see Tables 1 and 4). For example, exploration is a default behaviour (occurring when no
significant threats or opportunities are detected) which can reduce inherent uncertainty through
random movement that leads to the collection of information about the animal’s surroundings. A
reflex is a reactive behaviour designed to protect an animal or its resources from threats (e.g.,
startle), or to take advantage of an opportunity which provides immediate benefits (e.g., absorb
nutrients). Reflexes include the cervical contraction reflex in Mammals, and the suckle reflex in
infants. Defensive reflexes include the annelid escape reflex, the emetic reflex in a crocodile, pupil
dilation in response to light, and the human blink reflex.
Table 4: Classes of behaviour production units
CONTROL LEVEL PHYSIOLOGICAL SITUATIONAL APTITUDINAL
Reactive reflex instinct exploration
Motivated drive emotion play
Instincts are also reactive, but put an animal in a better position for survival or reproduction, without
a directly linked physiological benefit. They can be built from combinations of reflexes. For
example, fish can learn to associate cues of danger (e.g., conspecific injury signals) with predator
cues. (Brosnan, Earley & Dugatkin, 2003) They can also learn to associate habitat dimension cues
(e.g., water temperature) with presence of the predator (i.e., predator habitat recognition). Both of
these mechanisms lead to flight strategies (a predator escape reflex). In combination, they produce
an instinct to evade risky habitats, whether predators are present or not. (Kelley & Magurran, 2003)
Motivated behaviours include drives, which cause goal-directed sequences of behaviour which
result in the direct acquisition of fitness-related benefits.
Drives are therefore concerned with
achieving optimal levels of resources in an animal’s body or within an animal’s immediate control,
because these actions provide immediate evolutionary benefits. The regulation of the internal
milieu may involve consumption, or bringing external resources inside the body (hunger, thirst, or
acquiring male gametes, if female); however, sometimes it involves expelling excess resources
(e.g., waste products or, if male, gametes). What we call play is motivated behaviour with the
primary function of acquiring new knowledge or skills. As with exploration, play is not a response to
a particular indicator, but rather the lack of significant problems or opportunities; it is a default
behaviour for those with an abundance of energy and no other pressing needs (Burghardt, 2005).
Play not only results from the existence of surplus resources, it also creates new resources which
can be used later. Play may be associated with a physiological or situational end-state, but its
function is to enhance the organism’s aptitudes, which can then be subsequently employed in the
acquisition of evolutionary benefits. (Fagen, 1991)
Our definition of emotions as motivated behaviours which leave animals in situational end-states is
different from those in the social psychological literature, where emotions have been conceived as:
• interrupting behaviour by focusing or redirecting attention (Sloman, 2001) or limiting the
search for behavioural options (Evans, 2002)
• valuing events or possible behavioural options (emotion as any form of affect) (Dolan,
• finding the meaning of events (i.e., appraisal theory (Arnold, 1960; Lazarus, 1991; Scherer,
• helping people adjust to their local -- that is, cultural – circumstances (i.e., social
constructionism (Averill, 1980; Oatley, 1993))
• coordinating the physiological, expressive, subjective experiential and behavioural aspects
of responses within a person (Levenson, 1999)
• interpreting feelings (Clore, 1994; Damasio, 1994; James, 1884; Prinz, 2004)
• solving recurrent problems important to biological fitness (Delancey, 2002; Ekman, 1999;
Nesse, 1990; Tooby & Cosmides, 1990)
Our definition does bear a relationship to the last of these – the evolutionary psychological
definition, which is broader than ours. Nevertheless, many behaviours which would commonly be
considered emotional fit within our categorization. For example, care for family members (nurture
and pair-bond love), the quest for social status (ambition), group membership (belonging), territory
and money (resource acquisition/greed) all fit within our category of emotions.
Planning, however, is qualitatively different from these other classes of BPU. By planning we mean
goal-directed behaviour designed to reach arbitrarily distant or abstract end-states, or ‘objectives’.
(Fincham et al., 2002; van den Heuvel et al., 2005) Planning requires the ability to generate
representations of future action sequences and their likely outcomes. (Miller, Galanter & Pribram,
1960) By being able to represent goals explicitly as goals, end-states became a variable for mental
manipulation, or part of the process of behaviour production itself, rather than simply the
consequence of enacting a BPU. In effect, end-states were internalized, to be constructed, learned
and freely chosen. (Cosmides & Tooby, 2000; Stanovich, 2004) For this reason, planning can be
used to achieve end-states of any kind, including ones which cannot be specified before-hand.
Thus planning can serve physiological or situational ends, but can also be used to enhance
aptitudes. Indeed, planning can choose to pursue end-states which evolution would not select (i.e.,
maladaptive end-states). Thus, although the planning facility or BPU is an evolved structure, the
outcomes it produces are not necessarily consistent with evolutionary logic.
Despite the sophisticated nature of this complex cognition in higher Primates, including humans,
much of it goes on ‘implicitly’ in the sense of occurring below conscious awareness. (Bargh &
Chartrand, 1999; Reber, 1996; Wilson, 2004) The ability to manipulate mental representations
does not imply that one is consciously aware of one’s own mental states. For example, Primates
can engage in deception of conspecifics without necessarily congratulating themselves on their
cleverness. (Whiten & Byrne, 1991). We therefore argue that there are two systems for planning:
implicit and explicit. (Camerer, Loewenstein & Prelec, 2005; Rolls, 2005) Implicit cognition is
inaccessible to consciousness due to its sub-symbolic, distributed representation in the brain, while
explicit knowledge is captured (in computational models at least) by symbolic, localized
representations, in which each processing unit is more easily interpretable and has a clearer
conceptual meaning (Clark & Karmiloff-Smith, 1993; Sun, Slusarz & Terry, 2005) Different regions
of the brain appear to be involved in the two different kinds of systems. (Keele et al., 2003; Posner,
DiGirolamo & Fernandez-Duque, 1997) Brain imaging suggests the implicit evaluation system is
primarily located in the orbitofrontal and cingulate cortex, while the explicit system is located in the
lateral prefrontal and parietal cortex. (Rolls, 2005)
Formulating our classes of behaviour in this simple way allows us to identify relationships between
them. For example, instincts and emotions share the characteristic of leaving animals in situational
end-states, but differ from each other in their degree of control. We believe this is both a powerful
and parsimonious way of classifying behaviour that should have considerable utility in reducing the
confusion that surrounds claims about the psychological qualities of animal behaviour (e.g.,
whether seed-caching by birds is a form of planned behaviour (Raby et al., 2007; Suddendorf &
Thus far we have deduced how history and function, the two necessary characteristics of
evolutionary kinds, should shape the structures that produce behaviour. However, the types and
classes of BPUs we have identified thus far are still unlikely to be the proximal cause of behaviour
in a given situation. This is because animals recognize themselves to be in a particular context
which requires a particular kind of response. The ability to recognize the nature of different
situations and to select appropriate strategies must be the job of particular evolved mental
algorithms. For example, within the class of situations that require a motivated response to change
one’s relationship to the world (i.e., an emotional situation), there is still a considerable difference in
the kinds of computation, and hence behaviour, needed to repair one’s social status and to defend
one’s territory. Hence, we expect that natural selection will have devised BPUs able to respond
appropriately to situations of evolutionary significance which have been recurrent in an animal’s
phylogenetic history using algorithms specific to particular strategies. (Tooby & Cosmides, 1992)
These mechanisms for generating behaviour must be present in all the members of the species
which have faced similar ecological problems throughout their phylogenetic history. These
mechanisms must be BPUs (i.e., natural kinds), because one can extrapolate from their presence
in one member of a species to another. (Boyd, 1991; Millikan, in press)
Within the classes of BPUs we have discussed must therefore be ‘individual’ kinds designed to
address particular evolutionary tasks set animals by their ecological niche. For example, startle is a
reflex that helps people quickly redirect their attention in an appropriate direction. People also tend
to recoil from faeces, which is an instinct to avoid the environmental sources of infectious disease.
(Curtis, Aunger & Rabie, 2004) Defending offspring is an emotion because it conserves the
situation in which kin can serve as instruments for further gene reproduction. Pretend aggression is
a form of play because it is primarily concerned with acquiring skills at winning contested resources
or competing for social status. (Panksepp, 1998)
The number of individual kinds which can be defined in this way will be large, but not very large,
because the number of important evolutionary tasks is restricted to the dimensions of an animal’s
niche which distinguish it from that of other species in the same ecology. (Hutchinson, 1944) By the
same token, individual kinds will vary from species to species, as the niches of different species
cannot overlap completely. (Mayr, 1963)
A major exception to the argument that individual-level production mechanisms are BPUs is
executive behaviours. Although the brain-based systems which implement such behaviours are
themselves evolved, we do not believe that there are units within the neocortex dedicated to the
achievement of particular objectives. The plans which executive kinds produce are therefore not
evolved but constructed individually to solve perceived problems on the fly. Thus, planning is both
a class and an individual BPU: there is only one BPU capable of meta-representation and the
pursuit of objectives, most likely in the isomorphic prefrontal cortex.
In this paper, we have used deduction from the necessary characteristics of evolutionary kinds to
define natural kinds for producing behaviour at three levels. Our process of theory building by
deduction (albeit inspired by current empirical findings) follows a common pattern in science. For
example, the structure of atoms was predicted by Bohr before it was ever observed (Bohr, 1913);
the structure of genes was deduced from patterns observed in X-ray crystallographic plates
(Watson & Crick, 1953); and the means by which genes express themselves was inferred from
modelling prior to observation (Gamow, 1954). It is therefore reasonable to argue that deduction
should also play a major role in the identification of natural psychological kinds.
What we call ‘types’ of BPUs are defined with respect to the complexity of their physical
instantiation (i.e., implemented either as simple reactive BPUs, motivational BPUs of intermediate
complexity, or truly complex executive BPUs). It is also possible to define ‘classes’ of BPUs within
these types as a function of which kind of end-state such kinds produce (i.e., those providing
physiological, situational and aptitudinal benefits). Finally, within classes are ‘individual’ BPUs
which provide solutions to important evolutionary problems that living in particular kinds of
ecological niches has required animals to solve (e.g., hunger, thirst, get mate). If statistical
regularities in the world endure, they allow consistent selection pressures to operate over extended
periods of time, and thus produce evolved structures within organisms – our natural kinds.
Our types of BPUs (e.g., reactive, motivated) are physically similar (e.g., of simple or intermediate
complexity, respectively), although they may not overlap materially – they can be implemented by
different parts of the brain. In this, they resemble substances, a chemical natural kind, which are
composed of molecules, but not necessarily the same kinds of molecules (e.g., gold versus sodium
chloride). Our classes of BPUs (e.g., reflex, emotion) are linked by function and history, but not
necessarily structurally. This makes them similar to the biological kind of organs (e.g., heart, lung),
which share functional characteristics, but are morphologically and physically distinct. Individual
BPUs, our fundamental natural kind, are the equivalent of atoms or cells in these other sciences:
they are the basic physical building blocks of psychology, with an isolatable physical presence in
the brain; they are the way in which an individual brain physically produces an action, episode or
plan that produces some kind of evolutionary benefit. These parallels with the kinds of other
sciences are reassuring because they suggest we may have tapped into significant properties of
the multiple levels of natural kinds more generally.
We are not the first to hypothesize the existence of psychological kinds. Debates concerning
potential natural kinds in psychology have centred around cognition (Pylyshyn, 1984),
consciousness (Hardcastle, 1995), knowledge (Kornblith, 2002), concept (Machery, 2005),
psychiatric disorders (Zacher, 2000) particular emotions or the category of emotion (Barrett, 2006;
Charland, 2002; Griffiths, 1997; Griffiths, 2004a; Griffiths, 2004b), and human kinds (Cooper, 2004;
Hacking, 2002; Haslam, 1998). None of these proposals (except that for emotion) have met with
much acceptance, probably because no clear way has been found to distinguish different members
of the kind within these basic categories.
We are also not the first to postulate the existence of natural kinds related to behaviour. One
influential predecessor in this endeavour has been ethology. In the 20
century, this major effort
was devoted to the discovery of animal behavioural kinds with methods adopted from the natural
sciences. (Lorenz, 1950) However, ethology produced no widely accepted taxonomy of
behavioural kinds, even though numerous attempts have been made to construct schemes
applicable to any species – that is, a ‘standardized’ description or ‘ontology’ of animal behaviour.
(Catton et al., 2003; Golani, 1976; Schleidt et al., 1984); the Animal Behavior Ontology Project
[http://mlsource.ornith.cornell.edu/ethodata/]. We hope that the scheme provided here can provide
new direction to this project.
Our theory can be seen as a modification of expected utility theory in economics (Savage, 1954;
von Neumann & Morgenstern, 1944) or expectancy value models in psychology (Bandura, 1986;
Fishbein & Ajzen, 1975), where the value of a behaviour is determined by the odds the behaviour
will put an animal in a particular state times the value it attaches to being in that state. In our case,
value is measured not in terms of utility or perceived costs and benefits, but in terms of
evolutionary benefits. We argue that behaviour is valued by the probability it will put the animal in a
particular kind of end-state, influenced by the evolutionary value of that end-state (i.e., in terms of
its likely contribution to biological fitness). Our primary hypothesis is that we can categorize end-
states by the nature of their relationship to evolutionary benefits: direct provision, indirect provision
or even more indirect provision.
Of course behaving animals don’t explicitly calculate and compare expected evolutionary benefits
before acting. (Utility is a similar convenient fiction, and psychologists often assume decision-
making is only implicit as well.) The proximate measure of evolutionary benefits used by brains is
psychological reward: natural selection has tuned brains to like feedback from adaptive behaviour
and to dislike maladaptive behaviour. In this way, mechanisms that produce adaptive behaviour
are reinforced, so that biological fitness tends to increase via behaviour.
If we are right, and the BPUs we have deduced really serve as the fundamental building blocks of
psychology and behavioural science, then identifying behavioural production mechanisms should
give scientists a number of additional tools for empirical and theoretical work. We see the
perspective we have developed as having practical implications in four primary spheres: for the
neurosciences, behavioural sciences, and evolutionary biology, and with respect to interdisciplinary
The first domain of implication concerns the brain sciences. The project to investigate the structure
and function of the brain through neuroimaging badly needs a theoretical approach to behaviour. At
present neuroscientists use common sense and intuition to determine the stimuli for their
explorations of brain function using brain scanning technologies. These stimuli should rather be
chosen from principle. For example, to determine how disgust works in the brain, scientists should
first ascertain the purpose of the disgust BPU. Research suggests that disgust is a motivation to
avoid the substances and situations that would have caused disease in our ancestors. (Curtis et
al., 2004) Second, a set of stimuli should be found which, as far as is possible, mimic ancestral
disease threats. Third, subjects should be exposed to these stimuli in fMRI tests, with the results
being used to map the brain centres involved in disgust responses. The lack of principled sets of
stimuli, due to a lack of a proper definition of the relevant domain of behaviour, is likely to be one of
the reasons for continued controversy as to which parts of the brain are responsible for what
Further, we can predict just how particular BPUs are implemented in the brain based on our
argument about the order in which different classes of BPU originated. For example, being among
the oldest BPUs, gene-based developmental programmes will have had time to evolve in
association with particular reflexes, guaranteeing that they regularly appear as similar circuits in
different brains (to be found in older parts of the human brain). In particular, as short, dedicated
‘arcs’ of neurons, it is likely that there is little physical overlap in the circuitry used to implement
different reflexive BPUs. Thus, the members of a particular kind of behaviour such as the startle
reflex should reappear in similar fashion in different animals. Different reflexive BPUs (e.g., startle
and withdrawal) will also share a similar, evolved structure – the reflex arc. So we can identify both
a common history and form for the class of reflexive BPUs.
On the other hand, the planning BPU evolved relatively recently, and in humans is associated with
the prefrontal cortex. (Adolphs, 2001; Christoff & Gabrieli, 2000; Miller, Freedman & Wallis, 2002) It
is likely that the prefrontal cortex implements this executive BPU using information in distributed
networks spread widely through the prefrontal cortex, while involving older parts of the brain as
well. Therefore, the executive BPU may not be implemented in exactly the same way in the brains
of different individuals. The BPU for planning probably has a low level of genetic specificity, and is
implemented using functionally and morphologically similar tissue, the isocortex. (Fuster, 2003;
Mountcastle, 1998) So there is a lower likelihood that planning will be associated with
developmental programmes or even structurally distinct ones. (It is this generality of executive
function, and the lack of evolved end-states toward which they are directed, which enables
planning to address novel objectives without regard to specific goals.) We therefore expect the
executive BPU to be less ‘locatable’ and display more connectivity to memory and motivational
BPUs than reflexes (e.g., using brain imaging).
In between are the motivated BPUs. Although these are most likely modular, and so should have
some aspects which are dedicated to specific purposes, they are also likely to overlap in their
physical instantiation, with individual drives or emotions making use of components which serve as
reactive circuits under other circumstances. Since evolution tends to tinker with pre-existing forms,
they should invoke use of components that evolved earlier, and which can be shared between
BPUs. Determining whether these deductions about the physical instantiation of different types of
BPU are correct constitutes a rich and significant programme of research.
The second class of implications of our classification of behaviour concerns the study of behaviour
itself. Behavioural scientists (e.g., ecologists or ethologists) want to know the adaptive function of
particular behaviours; health psychologists and marketers want to better understand behaviour so
as to change it. However, we suspect that it will never be possible to determine the level of control
over individual behaviours. Observation of a single instance of any behaviour can be interpreted as
the result of the lowest level of control, as a behaviour built into an animal’s repertoire by genetic
evolution. Even complex sequences can be found to be instinctive rituals if they are repeated on
numerous similar occasions and are species-typical.
The behavioural scientist’s job, then, is to look for statistical regularities in behaviour with respect to
end-states. The stream of behaviour can be split into periods between physiological end-states
which provide evolutionary benefits (e.g., from copulation to sleep to eating to predator defence to
status improvement). Typically, the behavioural scientist will be looking for variation in the stream
of behaviour from one kind of end-state to another. Only through comparison with the behaviour of
the same animal or other animals in the species, resulting in the same category of beneficial end-
state, can it be inferred that behaviour has likely been reactive, motivated or planned. A good
indicator of the motivated nature of a stream of behaviour is that it exhibits a point at which some
evolutionary benefit within immediate reach has been ignored; an indicator of planned behaviour,
that some end-state achievable through motivated action has been ignored. Multiple routes which
reach the same category of end-state (e.g., copulation, fruit consumption, territory acquisition) will
be exhibited if behaviour is motivated. Plans will exhibit patterns in the particular sequences in
which beneficial end-states are reached. For example, plans that require high levels of skill to
achieve will tend to require that multiple episodes of playful behaviour occur before the objective is
achieved. Plans thus exhibit sequence dependence in motivated behaviours.
From the perspective of our own profession, health promotion, knowing which level of control is
responsible for a behaviour is fundamental (and the reason that we undertook this project)..
Recognizing that changing a behaviour will require overcoming motivation rather than reactive
control, for example, increases the tool kit available for changing that behaviour. Motivated
behaviour can be changed by providing incentives, knowledge, or physiological sensations,
whereas changing reactive behaviour can only be achieved through the provision of evolved cues.
Knowing which kind of end-state a behaviour results in can also have implications for the means
available to change it. Behaviours providing physiological benefits, even when unhealthy (e.g.,
drug-taking, sugar consumption), are more difficult to overcome than situational behaviours such
as not wearing seat belts, because immediate rewards have to be fought against.
Third, for evolutionary biologists, it should be possible to identify the behaviour kinds we have
established in the brains of related species in the human lineage. For example, both rat and human
brains should exhibit mental facilities to implement drives and emotions which are homologous
structures. This project is already underway, with a number of such homologies having been
discovered. (Panksepp, 1998; Streidter, 1998)
An empirical deduction from our approach is that animals in any given clade will only have certain
types of production unit available for their behavioural repertoire. Thus, Chordates prior to
Vertebrates will be ‘instinctual creatures’ (i.e., limited to instincts, reflexes and exploration), while
Mammals will also display drives, emotive and play behaviour, and planning will be limited to
Primates. These represent a few examples of the significant potential advances in the study of
behaviour which we hope will materialize through future studies of the natural kinds we have
We have also made specific claims about the order in which the different BPUs have evolved.
Primitive functions tend to arise earlier in development and to be more similar among species in a
clade (Kirschner & Gerhart, 2006; West-Eberhard, 2003). Hence among animals in the human
lineage, later-evolved BPUs should develop later in brain ontogeny, and diverge more greatly
between species in the human clade. We thus expect most animals will not exhibit higher cognitive
kinds, causing them to diverge from Primates, and that these cognitive kinds will only develop
relatively late in human ontogeny (i.e., with the frontal cortex, which develops last (Fuster, 2003)),
while earlier-evolved BPUs, such as instincts, will be found in earlier-developing parts of the human
Finally, we believe our approach has implications for interdisciplinary communication and
collaboration. We have argued that psychology cannot advance as a science without terms that
can attract agreement within and between the behavioural sciences. Our argument is that the best
way to do this is to equate psychological constructs to biological natural kinds. Such kinds can form
the foundation for interdisciplinary cooperation on the study of behaviour. It is unfortunate, for
example, that a term as basic to brain and behaviour as ‘emotion’ still has no agreed-upon
definition. Our approach provides a short crisp definition, not from introspection or empirical
observation, but from the perspective of function and evolutionary history. Emotion is a class of
BPUs in the brain that motivate behaviour designed to improve an animal’s situation with respect to
its social world.
We can also compare predictions made from our approach about the nature of individual BPUs
with those derived from a major alternative: folk psychology. Although we use non-standard means
to classify behaviour into kinds, our typology does approximately map onto intuitive or ‘folk’
categorizations of behaviour in many cases. Hence, hunger is a drive, not because it causes a
particular set of physiological states or feelings, but because it leads to evolutionarily significant
returns through flexibly controlled, motivated consumption. Similarly, the emotion of pair-bond love
leads an animal to work; it risks physical harm to defend rights of access to, or control of, a mate (a
situational end-state). Loving behaviours can take many forms, and may endure until a particular
strategic relationship is achieved. These two conditions define pair-bond love as an emotion.
However, a counter-intuitive deduction from our approach is that behaviours commonly called
‘disgust’ and ‘fear’ should often be considered drives rather than emotions. Fleeing from predation
is an effort to avoid one’s bodily resources becoming the predator’s resources. Similarly,
behaviours such as walking around excreta or rotten meat, or shunning someone who appears to
be ill, involve a ‘negative appetite’ for avoiding resources crossing the body boundary (in this case
wanting to avoid being eaten from inside by pathogens). Our presumption is that these behaviours
don’t involve the same complex meta-representation of prey or parasite as social interactions, and
so are not emotional in nature. Emotional fear we reserve for behaviours which avoid conspecifics
as threats to body or resources; emotional disgust is restricted to those motivated and contingent
(i.e., strategic) behaviours which serve to ostracize, or shun, social ‘parasites’ (Curtis & Biran,
2001) Thus the overlap between our categories of behaviour and those given common-sensical
names is not perfect. Such differences can be used to make unique predictions that can support or
refute our approach.
There is hardly a more powerful tool in science than what philosophers call ‘natural kinds’. (Wagner
& Wagner, 2003) Natural kinds are forms ‘given by nature’, not categories artificially constructed by
the human mind (Boyd, 1991; Millikan, 1984). They are structures generated by processes that
have distinctive intrinsic natures described by the causal factors at work in their production and
maintenance. In this paper we have argued that natural kinds for psychology can be found by
dividing behaviour into categories based on the kinds of behaviour production units which caused
them. In particular, we have made use of a three-part division of the degree of psychological
control over behaviour execution and an original distinction we have made between kinds of
evolutionary end-states into which animals are put by this behaviour, to identify these kinds. We
argue that the selective forces producing new levels of sophistication in the production of behaviour
were the ability to control behaviour over longer periods, thus obtaining higher average benefits,
and given a level of control, by the ability to devise more and more indirect means of acquiring
evolutionary benefits by reaching different kinds of end-states. While our account necessarily
remains only plausible at present, we believe that it is consistent with a wide variety of evidence.
We therefore believe this story lends support to our contention that natural kinds in animal
behaviour exist, and have evolved in an understandable, orderly sequence.
We believe our approach has three major advantages over previous efforts to define classes of
behaviour. First, it is uniquely constrained by both functional and historical considerations to tell a
plausible story about how the various kinds of production first evolved, from one level to the next,
with support from known developments in nervous systems, learning and animal task repertoires.
By basing our approach in evolutionary theory, with particular reference to major transitions in
complexity, we can make explicit claims about the order, timing, and means by which each kind of
cause of behaviour arose. Second, our account combines production types and output types (end-
states) into a comprehensive and interdependent model of the causes of behaviour; previous
models have been limited to one or the other of these two dimensions. Third, we have provided
new criteria for rigorously distinguishing among the various kinds of behaviour production systems.
These criteria result in new claims about the functions of instincts and emotions, for example.
By redefining reflexes, instincts, exploration, drives, emotions, play and plans according to their
evolved functions rather than by their mechanics or by the subjective states they create, we hope
to provide a principled vocabulary that can be shared by behavioural scientists in any discipline,
and applied to any species in our lineage. Most of these terms have long histories in psychology, of
course, and have been abandoned by many precisely because of their imprecision. However, we
believe that these concepts will continue to have value once couched in a natural kinds framework.
Thanks to Adam Biran, Kalina Christoff, Barbara Findlay, Carlos Gershenson, Ara Norenzayan,
John Odling-Smee, Miguel Rubio-Godoy, Thomas Reydon, Beth Scott, Szymon Wichary and
several anonymous reviewers for comments on earlier versions.
Adolphs, R. (2001). The neurobiology of social cognition. Current Opinion in Neurobiology 11, 231-
Allen, C. & Bekoff, M. (1997). Species of Mind: The Philosophy and Biology of Cognitive Ethology.
MIT Press, Cambridge, MA.
Allman, J. M. (1999). Evolving Brains. Scientific American Library, New York.
Amabile, T. M. (1996). Creativity in Context. Westview Press, Boulder, Co.
Arnold, M. B. (1960). Emotion and Personality. Columbia University Press, New York.
Austin, J. T. & Vancouver, J. B. (1996). Goal constructs in psychology: Structure, process, and
content. Psychological Bulletin 120, 338-375.
Averill, J. R. (1980). A constructionist view of emotion. In Emotion: Theory, Research, and
Experience (ed. R. Plutchik and H. Kellerman), pp. 305-339. Academic Press, New York.
Bandura, A. (1986). Social Foundations of Thought and Action: A Social Cognitive Theory. Prentice
Hall, Englewood Cliffs, NJ.
Bargh, J. A. & Chartrand, T. L. (1999). The unbearable automaticity of being. American
Psychologist 54, 462-479.
Barker, R. G. & Wright, H. F. (1954). Midwest and its Children: The Psychological Ecology of an
American Town. Row, Peterson and Company, Evanston, Ill.
Barlow, H. B. (1994). What is the computational goal of the neocortex? In Large-scale Neuronal
Theories of the Brain (ed. C. Koch and J. L. Davis), pp. 1-22. MIT Press, Boston, MA.
Barrett, L. F. (2006). Emotions as natural kinds? Perspectives on Psychological Science 1, 28-58.
Berridge, K. C. (2004). Motivation concepts in behavioral neuroscience. Physiology and Behavior
Bohr, N. (1913). On the constitution of atoms and molecules. Philosophical Magazine 26, 1-25.
Bolker, J. A. (2000). Modularity in development and why it matters to evo-devo. American Zoologist
Boyd, R. N. (1991). Realism, anti-foundationalism and the enthusiasm for natural kinds.
Philosophical Studies 61, 127-148.
Boyd, R. N. (1999a). Kinds, complexity and multiple realization. Philosophical Studies 95, 67-98.
Boyd, R. N. (1999b). Homeostasis, species, and higher taxa. In Species: New Interdisciplinary
Essays (ed. R. A. Wilson), pp. 141-185. MIT Press, Cambridge, MA.
Brosnan, S., Earley, R. & Dugatkin, L. A. (2003). Observational learning and predator inspection in
guppies. Ethology 109, 823-833.
Burghardt, G. M. (2005). The Genesis of Animal Play. MIT Press, Cambridge, MA.
Butler, A. B. & Hodos, W. (2005). Comparative Vertebrate Neuroanatomy: Evolution and
Adaptation. 2nd ed. Wiley-Liss, New York.
Camerer, C., Loewenstein, G. & Prelec, D. (2005). Neuroeconomics: How neuroscience can inform
economics. Journal of Economic Literature 43, 9-64.
Catton, C., Wilson, C., Dalton, R. & Shotton, D. M. (2003). SABO: a Standard Animal Behaviour
Ontology (alpha version 0.1). vol. 2006.
Charland, L. C. (2002). The natural kind status of emotion. British Journal for the Philosophy of
Science 53, 511-537.
Christoff, K. & Gabrieli, J. D. E. (2000). The Frontopolar Cortex and Human Cognition: Evidence for
a Rostrocaudal Hierarchical Organization within the Human Prefrontal Cortex.
Psychobiology 28, 168-186.
Churchland, P. & Sejnowski, T. J. (1992). The Computational Brain. MIT Press, Boston, MA.
Clark, A. & Karmiloff-Smith, A. (1993). The Cognizer’s innards: A psychological and philosophical
perspective on the development of thought. Mind and Language 8, 487–519.
Clore, G. (1994). Why emotions are felt. In The Nature of Emotion: Fundamental Questions (ed. P.
Ekman and R. J. Davidson), pp. 103-111. Cambridge University Press, New York.
Coates, M., Jeffrey, J. & Ruta, M. (2002). Fins to limbs: what the fossils say. Evolution and
Development 4, 390–401.
Cooper, R. (2004). Why Hacking is wrong about human kinds. British Journal for the Philosophy of
Science 55, 73-85.
Cosmides, L. & Tooby, J. (2000). Consider the source: The evolution of adaptations for decoupling
and metarepresentation. In Metarepresentation: A Multidisciplinary Perspective (ed. D.
Sperber). Oxford University Press, New York.
Curtis, V., Aunger, R. & Rabie, T. (2004). Evidence that disgust evolved to protect from risk of
disease. Proceedings: Biological Sciences (Biology Letters) 277, 131-133.
Curtis, V. & Biran, A. (2001). Dirt, disgust and disease. Perspectives in Biology and Medicine 44,
Cziko, G. (1995). Without Miracles: Universal Selection Theory and the Second Darwinian
Revolution. MIT Press, Cambridge, MA.
Damasio, A. R. (1994). Descartes' Error: Emotion, Reason, and the Human Brain. G. P. Putnam's
Sons, New York.
Damasio, A. R. (2003). Looking for Spinoza. Harcourt, New York.
Dawkins, R. (1976). Hierarchical organisation: a candidate principle for ethology. In Growing Points
In Ethology (ed. P. Bateson and R. Hinde), pp. 7-54. Cambridge University Press,
Dayan, P. & Abbott, L. F. (2001). Theoretical Neuroscience: Computational and Mathematical
Modeling of Neural Systems. MIT Press, Boston, MA.
Deci, E. L. (1971). Effects of externally mediated rewards on intrinsic motivation. Journal of
Personality and Social Psychology 18, 105-115.
Deci, E. L. (1975). Intrinsic Motivation. Plenum Publishing Co., New York.
Deci, E. L. & Ryan, R. M. (1985a). Intrinsic Motivation and Self-Determination in Human Behavior.
Plenum, New York.
Deci, E. L. & Ryan, R. M. (1985b). Intrinsic Motivation and Self-Determination in Human
Behavior. Plenum Press, New York.
Delancey, C. (2002). Passionate Engines: What Emotions Reveal about Mind and Artificial
Intelligence. Oxford University Press, New York.
Dennett, D. (1996). Kinds of Minds: Toward an Understanding of Consciousness. Basic Books,
Dienes, Z. & Perner, J. (1999). A theory of implicit and explicit knowledge. Behavioral and Brain
Sciences 22, 735-808.
Dolan, R. J. (2002). Emotion,cognition,and behavior. Science 298, 1191-1194.
Druce, M. R., Small, C. J. & Bloom, S. R. (2005). Minireview: Gut peptides regulating satiety.
Endocrinology 145, 2660-2665.
Dupré, J. A. (2000). Natural kinds. In A Companion to the Philosophy of Science (ed. W. H.
Newton-Smith), pp. 311-319. Blackwell, Oxford.
Editors. (2006). The American Heritage Dictionary of the English Language, Fourth Edition.
Houghton Mifflin, New York.
Ekman, P. (1999). Basic emotion. In Handbook of Cognition and Emotion (ed. T. Dalgleish and M.
Power), pp. 45-60. John Wiley & Sons, Ltd., Sussex, U.K.
Evans, D. (2002). The search hypothesis of emotion. British Journal for the Philosophy of Science
Fagen, R. (1991). Animal Play Behavior. Oxford University Press, Oxford.
Fernald, R. D. (1997). The evolution of eyes. Brain Behavior and Evolution 50, 253-259.
Fincham, J. M., Carter, C. S., van Veen, V., Stenger, V. A. & Anderson, J. R. (2002). Neural
mechanisms of planning: A computational analysis using event-related fMRI. Proc.
National Academy of Sciences 99, 3346-3351.
Fishbein, M. & Ajzen, I. (1975). Belief, attitude, intention, and behavior: an introduction to theory
and research. Addison–Wesley, Reading, MA.
Franken, R. (2001). Human Motivation; 5th ed. Brooks/Cole, Pacific Grove, CA.
Frederich, S., Lowenstein, G. & O’Donoghue, T. (2002). Time discounting and time preference: A
critical review. Journal of Economic Literature 40, 351-401.
Freeman, W. J. (1999). How the Brain Makes up its Mind. Weidenfeld and Nicholson, London.
Freeman, W. J. (2000). Emotion is essential to all intentional behaviors. In Emotion, Development,
and Self-Organization: Dynamic Systems Approaches to Emotional Development (ed. M.
D. Lewis and I. Granic), pp. 209-235. Cambridge University Press, Cambridge.
Fuster, J. (2003). Cortex and Mind: Unifying Cognition. Oxford University Press, Oxford.
Gamow, G. (1954). Possible relation between deoxyribonucleic acid and protein structures. Nature
Geary, D. C. & Huffman, K. J. (2002). Brain and cognitive evolution: Forms of modularity and
functions of mind. Psychological Bulletin 128, 667-698.
Gehring, W. J. & Ikeo, K. (1999). Pax 6: Mastering eye morphogenesis and eye evolution. Trends
in Genetics 15, 371-377.
Gershenson, C. (2001). Artificial Societies of Intelligent Agents. Ms Thesis thesis, Fundación Arturo
Ghiselin, M. T. (1974). A radical solution to the species problem. Systematic Zoology 23, 536-544.
Gibbs, J., Maddison, S. P. & Rolls, E. T. (1981). The satiety role of the small intestine in sham
feeding rhesus monkeys. Journal of Comparative and Physiological Psychology 95, 1003-
Godfrey-Smith, P. (1996). Complexity and the Function of Mind in Nature. Cambridge University
Golani, I. (1976). Homeostatic motor processes in mammalian interactions: a choreography
of display. In Perspectives in Ethology, vol. 2 (ed. P. P. G. Bateson and P. H. Klopfer), pp. 69-134.
Plenum, New York.
Gray, J. R., Braver, T. S. & Raichle, M. E. (2002). Integration of emotion and cognition in the lateral
prefrontal cortex. Proceedings of the National Academy of Sciences, USA 99, 4115-4120.
Griffiths, P. (1997). What Emotions Really Are: The Problem of Psychological Categories.
University of Chicago Press, Chicago.
Griffiths, P. E. (1999). Squaring the circle: natural kinds with historical essences. In Species: New
Interdisciplinary Essays (ed. R. A. Wilson), pp. 209-228. MIT Press, Cambridge, MA.
Griffiths, P. E. (2004a). Is emotion a natural kind? In Thinking About Feeling: Contemporary
Philosophers on Emotions (ed. R. C. Solomon), pp. 233-249. Oxford University Press, New
Griffiths, P. E. (2004b). Emotions as natural and normative kinds. Philosophy of Science 71, 901-
Hacking, I. (2002). Making up people. In Historical Ontology (ed. I. Hacking), pp. 99-114. Harvard
University Press, Cambridge, MA.
Hardcastle, V. G. (1995). Locating Consciousness. John Benjamins, Amsterdam & Philadelphia.
Harnish, R. (2001). Minds, Brains, Computers: An Historical Introduction to the Foundations of
Cognitive Science. Blackwell, London.
Hartwell, L., H. , Hopfield, J. J., Leibler, S. & Murray, A. W. (1999). From molecular to modular cell
biology. Nature 402, C47–C52.
Haslam, N. (1998). Natural kinds, human kinds, and essentialism. Social Research 65, 291-314.
Haslam, N. (2002). Natural kinds, practical kinds and psychiatric categories: Review of Zachar on
Biological-Psychiatry. Psycoloquy 13, Biological Psychiatry (2).
Hawkins, J. & Blakeslee, S. (2004). On Intelligence. Henry Holt, New York.
Hebb, D. O. (1949). The Organization of Behavior. John Wiley and Sons, New York.
Hennig, W. (1966). Phylogenetic Systematics. University of Illinois Press, Urbana.
Hughlings Jackson, J. (1958). Selected Writings of John Hughlings Jackson (Vols. 1 and 2).
Staples Press, London.
Hull, C. L. (1943). Principles of Behavior: An Introduction to Behavior Theory. Appleton-Century
Co., New York.
Hull, D. (1978). A matter of individuality. Philosophy of Science 45, 335-360.
Hutchinson, G. E. (1944). Limnological studies in Connecticut. VII. A critical examination of the
supposed relationship between phytoplankton periodicity and chemical changes in lake
waters. Ecology 26, 3-26.
James, W. (1884). What is an emotion? Mind 19, 188-204.
Johnson-Laird, P. N. (2006). How We Reason. Oxford University Press, Oxford.
Keele, S., Ivry, R., Mayr, U., Hazeltine, E. & Heuer, H. (2003). The cognitive and neural
architecture of sequence representation. Psychological Review 110, 316-339.
Keller, R. A., Boyd, R. N. & Wheeler, Q. D. (2003). The illogical basis of phylogenetic
nomenclature. Botanical Review 69, 93-110.
Kelley, J. L. & Magurran, A. E. (2003). Learned predator recognition and antipredator responses in
fish. Fish and Fisheries 4, 216-226.
Kelso, J. A. S. (1995). Dynamic Patterns: The Self-Organisation of Brain and Behavior. MIT Press,
Kirschner, M. W. & Gerhart, J. C. (2006). The Plausibility of Life. Yale University Press, New
Koechlin, E., Ody, C. & Kouneiher, F. (2003). The architecture of cognitive control in the human
prefrontal cortex. Science 302, 1181-1185.
Kornblith, H. (2002). Knowledge and Its Place in Nature. Clarendon Press, Oxford.
Kripke, S. A. (1972). Naming and Necessity. Harvard University Press, Cambridge, MA.
LaPorte, J. (2004). Natural Kinds and Conceptual Change. Cambridge University Press,
Lazarus, R. (1991). Emotion and Adaptation. Oxford University Press, New York.
LeDoux, J. E. (2000). Emotion circuits in the brain. Annual Review of Neuroscience 23, 155-184.
Lepper, M. R., Greene, D. & Nisbett, R. E. (1973). Undermining children's intrinsic interest with
extrinsic rewards: A test of the 'overjustification' hypothesis. Journal of Personality and
Social Psychology 28, 129-137.
Lepper, M. R. & Henderlong, J. (2000). Turning 'play' into 'work' and 'work' into 'play': 25 years of
research on intrinsic versus extrinsic motivation. In Intrinsic and Extrinsic Motivation: The
Search for optimal motivation and performance (ed. C. Sansone and J. M. Harackiewicz),
pp. 257-307. Academic, San Diego, CA.
Levenson, R. W. (1999). The intrapersonal functions of emotion. Cognition and Emotion 13, 481-
Lorenz, K. (1950). The comparative method in studying innate behaviour patterns. In Physiological
Mechanisms in Animal Behaviour (ed. J. F. Danielli and R. Brown). Cambridge University
Lorenz, K. (1965). Evolution and the Modification of Behavior. University of Chicago Press,
Machery, E. (2005). Concepts are not a natural kind. Philosophy of Science 72, 444-467.
Mallon, R. (2003). Social construction, social roles, and stability. In Socializing Metaphysics (ed. F.
Schmitt), pp. 327-353. Rowman and Littlefield, New York.
Maslow, A. H. (1943). A theory of human motivation. Psychological Review 50, 370-396.
Mayr, E. (1963). Animal Species and Evolution. Harvard University Press, Cambridge, MA.
McClure, S. M., Ericson, K. M., Laibson, D. I., Loewenstein, G. & Cohen, J. D. (2007). Time
discounting for primary rewards. Journal of Neuroscience 27, 5796-5804.
McClure, S. M., Laibson, D. I., Loewenstein, G. & Cohen, J. D. (2004). Separate neural systems
value immediate and delayed monetary rewards. Science 306, 503-507.
Mesulam, M.-M. (1998). From sensation to cognition. Brain 121, 1013-1052.
Michaud, F. (2002). EMIB -- Computational architecture based on emotion and motivation for
intentional selection and configuration of behaviour-producing modules. Cognitive Science
Quarterly 3-4, 340-361.
Miller, E. K. & Cohen, J. D. (2001). An integrative theory of prefrontal cortex function. Annual
Review of Neuroscience 24, 167-202.
Miller, E. K., Freedman, D. J. & Wallis, J. D. (2002). The prefrontal cortex: categories, concepts,
and cognition. Philosophical Transactions: Biological Sciences 357, 1123-1136.
Miller, G. A., Galanter, E. H. & Pribram, K. H. (1960). Plans and the Structure of Behaviour. Holt
Rinehart and Winston, New York.
Millikan, R. G. (1984). Language, Thought and Other Biological Categories. MIT Press, Cambridge,
Millikan, R. G. (1989). In defense of proper functions. Philosophy of Science 56, 288-302.
Millikan, R. G. (1999). Historical kinds and the "special sciences". Philosophical Studies 95, 45-65.
Millikan, R. G. (2000). On Clear and Confused Ideas: An Essay about Substance Concepts.
Cambridge University Press, Cambridge.
Millikan, R. G. (in press). Thoughts of real kinds. In The Oxford Handbook of Philosophy of
Psychology (ed. J. Prinz). Oxford University Press, Oxford.
Mountcastle, V. (1957). Modality and topographic properties of single neurons of cat’s somatic
sensory cortex. Journal of Neurophysiology 20, 408–434.
Mountcastle, V. B. (1998). Perceptual Neuroscience: The Cerebral Cortex. Harvard University
Press, Cambridge, MA.
Murray, H. A. (1938). Explorations in Personality. Oxford University Press, New York.
Myers, D. G. (2006). Psychology (8th ed). Worth, New York.
Nesse, R. M. (1990). Evolutionary explanations of emotion. Human Nature 1, 261-289.
Newell, A. (1990). Unified Theories of Cognition. Harvard University Press, Cambridge, MA.
Northcutt, R. G. (2002). Understanding vertebrate brain evolution. Integrative and Comparative
Biology 42, 743-756.
Northcutt, R. G. & Kaas, J. H. (1995). The emergence and evolution of mammalian neocortex.
Trends in the Neurosciences 18, 373-379.
Oatley, K. J. (1993). Social construction in emotion. In Handbook of Emotions (ed. M. Lewis and J.
Haviland), pp. 342-352. Guilford Press, New York.
Ochsner, K. N. & Gross, J. J. (2005). The cognitive control of emotion. Trends in Cognitive
Odling-Smee, F. J., Laland, K. N. & Feldman, M. (2003). Niche Construction: The Neglected
Process in Evolution. Princeton University Press, Princeton.
Olds, J. & Milner, P. M. (1954). Positive reinforcement produced by electrical stimulation of the
septal area and other regions of the rat brain. Journal of Comparative and Physiological
Psychology 47, 419-427.
Ortony, A., Norman, D. A. & Revelle, W. (2004). Affect and proto-affect in effective functioning. In
Who Needs Emotions? The Brain Meets the Robot (ed. J. M. Fellous and M. A. Arbib), pp.
173-202. Oxford University Press, Oxford.
Panksepp, J. (1998). Affective Neuroscience: The Foundations of Human and Animal Emotions.
Oxford University Press, New York.
Perner, J. (1991). Understanding the Representational Mind. MIT Press, Cambridge, MA.
Posner, M. I., DiGirolamo, G. J. & Fernandez-Duque, D. (1997). Brain mechanisms of cognitive
skills. Consciousness and Cognition 6, 267-290.
Prinz, J. (2004). Gut Reactions: A Perceptual Theory of Emotion. Oxford University Press, Oxford:.
Proust, J. (in press). Metacognition and animal rationality. In Rational Animals? (ed. S. Hurley and
M. Nudds). Oxford University Press, Oxford.
Putnam, H. (1975). The Meaning of ‘Meaning. In Mind, Language, and Reality (ed. H. Putnam), pp.
215-271. Cambridge University Press, Cambridge.
Pylyshyn, Z. W. (1984). Computation and Cognition: Toward a Foundation for Cognitive Science.
MIT Press, Cambridge, MA.
Quartz, S. (2001). Toward a developmental evolutionary psychology: Genes, development, and the
evolution of human cognitive architecture. In Evolutionary Psychology: Alternative
approaches (ed. S. Scher, M. Rauscher and (Eds).). Kluwer.
Quine, W. V. (1969). Natural kinds. In Essays in Honor of Carl. G. Hempel (ed. N. Rescher), pp. 5-
23. Reidel, Dordrecht.
Raby, C., Alexis, D., Dickinson, A. & Clayton, N. (2007). Planning for the future by western scrub-
jays. Nature 445, 919-21.
Raff, R. A. (1996). The Shape of Life. Chicago University Press, Chicago.
Reber, A. (1996). Implicit Learning and Tacit Knowledge: An Essay on the Cognitive Unconscious.
Oxford University Press, Oxford.
Rieppel, O. (2005a). Modules, kinds, and homology. Journal of Experimental Zoology B (Molecular
and Developmental Evolution) 304, 18-27.
Rieppel, O. (2005b). Monophyly, paraphyly, and natural kinds. Biology and Philosophy 20, 465-
Rolls, E. T. (1999). The Brain and Emotion. Oxford University Press, Oxford.
Rolls, E. T. (2005). Emotion Explained. Oxford University Press, Oxford.
Roseman, I. J. (1984). Cognitive determinants of emotion: A structural theory. In Review of
Personality and Social Psychology, vol. 5 (ed. P. Shaver), pp. 11-36. Sage, Beverly Hills,
Rosenthal, D. M. (2000). Consciousness and metacognition. In Metarepresentations: A
Multidisciplinary Perspective (ed. D. Sperber). Basil Blackwell, London.
Routtenberg, A. & Lindy, J. (1965). Effects of the availability of rewarding septal and hypothalamic
stimulation on bar pressing for food under conditions of deprivation. Journal of
Comparative and Physiological Psychology 60, 158-161.
Rubenstein, J. L., Anderson, S., Shi, L., Miyashita, L. E., Bulfone, A. & Hevner, R. (1999). Genetic
control of cotrtical regionalization and connectivity. Cerebral Cortex 9, 524-532.
Ryan, R. M. & Deci, E. L. (2000). Intrinsic and extrinsic motivations: Classic definitions and new
directions. Contemporary Educational Psychology 25, 54-67.
Savage, L. J. (1954). Foundations of Statistics. Dover, New York.
Schank, R. & Abelson, R. (1977). Scripts, Plans Goals and Understanding: An Inquiry Into Human
Knowledge Structures. Erlbaum, Hillsdale, NJ.
Scherer, K. R. (1984). On the nature and function of emotion: a component process approach. In
Approaches to Emotion (ed. K. R. Scherer and P. Ekman), pp. 293–317. Erlbaum,
Schleidt, W. M., Yakalis, G., Donnelly, M. & McGarry, J. (1984). A proposal for a standard
ethogram, exemplified by an ethogram of the bluebreasted quail (Coturnix chinensis).
Zeitschreift für Tierpsychologie 64, 193-220.
Sheffield, F. D. & Roby, T. B. (1950). Reward value of a non-nutritive sweet taste. Journal of
Comparative and Physiological Psychology 43, 471-481.
Skinner, B. F. (1938). The Behavior of Organisms: An Experimental Analysis. Prentice Hall,
Englewood Cliffs, NJ.
Sloman, A. (2001). Beyond shallow models of emotion. Cognitive Processing: International
Quarterly of Cognitive Science 2, 177-198.
Sperber, D. (2000). Metarepresentations in evolutionary perspective. In Metarepresentations: A
Multidisciplinary Perspective (ed. D. Sperber), pp. 117-137. Oxford University Press,
Stanovich, K. E. (2004). The Robot's Rebellion: Finding Meaning in the Age of Darwin. University
of Chicago Press, Chicago.
Sterelny, K. (1990). The Representational Theory of Mind: An Introduction. Basil Blackwell, Oxford.
Sterelny, K. (1998). Intentional agency and the metarepresentation hypothesis. Mind and
Language 13, 11-28.
Sterelny, K. (2003). Mind in a Hostile World. Oxford University Press, Oxford.
Streidter, G. F. (1998). Progress in the study of brain evolution: From speculative theories to
testable hypotheses. The Anatomical Record 253, 105 - 112.
Streidter, G. F. (2005). The Principles of Brain Evolution. Sinauer Associates, Sunderland, MA.
Suddendorf, T. & Busby, J. (2003). Mental time travel in animals? Trends in Cognitive Sciences 7,
Sun, R., Slusarz, P. & Terry, C. (2005). The interaction of the explicit and the implicit in skill
learning: A dual-process approach. Psychological Review 112, 159-192.
Swanson, L. W. (2003). Brain Architecture: Understanding the Basic Plan. Oxford University Press,
Thelen, E. & Smith, L. B. (1994). A Dynamical Systems Approach to the Development of Cognition
and Action. MIT Press, Cambridge, MA.
Tinbergen, N. (1951). The Study of Instinct. Clarendon Press, Oxford.
Tooby, J. & Cosmides, L. (1990). The past explains the present: Emotional adaptations and the
structure of ancestral environments. Ethology and Sociobiology 11, 375-424.
Tooby, J. & Cosmides, L. (1992). The psychological foundations of culture. In The Adapted Mind:
Evolutionary Psychology and the Generation of Culture (ed. J. H. Barkow, L. Cosmides
and J. Tooby), pp. 19-136. Oxford University Press, Oxford.
van den Heuvel, O., Veltman, D. J., Groenewegen, H. J., Cath, D. C., van Balkom, A. J. L. M., van
Hartskamp, J., Barkhof, F. & van Dyck, R. (2005). Frontal-striatal dysfunction during
planning in obsessive-compulsive disorder. Archives of General Psychiatry 62, 301-309.
von Neumann, J. & Morgenstern, O. (1944). Theory of Games and Economic Behavior. Princeton
University Press, Princeton.
Wagner, G. P. (1996). Homologues, natural kinds and the evolution of modularity. American
Zoologist 36, 36-43.
Wagner, W. & Wagner, G. P. (2003). Examining the modularity concept in evolutionary psychology:
The levels of genes, mind and culture. Journal of Cultural and Evolutionary Psychology 1.
Watson, J. D. & Crick, F. H. C. (1953). A structure for deoxyribose nucleic acid. Nature 171, 737-
West-Eberhard, M. J. (2003). Developmental Plasticity and Evolution. Oxford University Press,
White, R. W. (1959). Motivation reconsidered: The concept of competence. Psychological Review
Whiten, A. & Byrne, R. W. (1991). The emergence of metarepresentation in human ontogeny and
primate phylogeny. In Natural Theories of Mind (ed. W. A.). Basil Blackwell, London.
Wilson, T. D. (2004). Strangers to Ourselves: Discovering the Adaptive Unconscious. Belknap
Press, Cambridge, MA.
Wong, R. (2000). Motivation: A Biobehavioural Approach. Cambridge University Press, Cambridge.
Wood, J. N. & Grafman, J. (2003). Human prefrontal cortex: processing and representational
perspectives. Nature Reviews Neuroscience 4, 139-146.
Zacher, P. (2000). Psychological Concepts and Biological Psychiatry: A Philosophical Analysis.
John Benjamins, New York.
Zacks, J. M. & Tversky, B. (2001). Event structure in perception and conception. Psychological
Bulletin 127, 3-21.
Movement can be of the whole body, of body parts (e.g., swimming, tail-flicking, flint-knapping), or
ejections of matter or energy into the environment from inside the body (e.g., defecation, speech).
Behaviour must be self-powered (i.e., locomotory) rather than dependant on external forces (e.g.,
drifting in a current of water or air).
We do not count physiological movements such as heart-beats
or peristalsis as behaviour because they do not interact with the environment. Such internal
movements are also controlled by the autonomous nervous system rather than the central nervous
system, and so are independently regulated.
The behaviour production unit lets animals turn salient external and internal stimuli into adaptive
outputs or behaviour. We have thus set aside input side topics such as sensation, perception,
categorization and concept discrimination. We similarly also avoid issues such as selection among
competing behavioural options and motor control on the output side. Finally, we have set aside, at
least for the moment, some of the special complexities of human behaviour -- difficult issues such
as temperament, mood, expression, and the role of culture. For the most part, these either
modulate processes we do discuss or are late additions to behaviour production units.
Since behaviour does not fossilise, we cannot reconstruct the behavioural abilities of our
ancestors directly; we can only make inferences based on extant animals that may, or may not, be
representative of the behavioural capabilities at a certain point in our evolutionary past, given that
contemporary species have had many millions of years (in some cases) to evolve new abilities.
Our argument regarding the timing of new types of production systems is based not on the most
advanced contemporary exemplar in some clade, but rather on the inferred qualities of a
prototypical ancestral species in that group.
Just about any behaviour is likely to be accompanied by learning. The knowledge acquired in this
way can modify the way in which a particular BPU works in future. However, this does not make all
behaviour aptitudinal in our sense because the primary function of those behaviours is to achieve
other kinds of end-states, rather than being devoted to learning per se.
Experimental psychologists nowadays avoid use of the word ‘drive’, largely because a number of
results have been interpreted as refuting Hull’s classic drive reduction theory (Hull, 1943).
However, in each case, these experimental results can be interpreted as ‘tricking’ the reward
system in a way which natural selection cannot be expected to have foreseen. First, rats will work
to obtain saccharine rewards even when not calorie deficient (i.e., not in a state of ‘hunger’)
(Sheffield & Roby, 1950). However, saccharine constitutes an artificial, non-nutritive source of
sweet taste, which is a proximal cue of ripe fruit in an animal’s evolved niche. Second, hungry rats
will work to obtain direct electrical stimulation of their mesotelencephalic dopamine system, even
over the option to receive food (Olds & Milner, 1954; Routtenberg & Lindy, 1965). This surgical
procedure short-circuits the evolved reward system; selection is unlikely to predict a technology
which provides rewards in the absence of the requisite biological resources following behaviour.
Showing that maladaptive outcomes can be obtained through such artificial procedures does not
refute the validity of the general claim that drives have evolved to produce need-directed behaviour
which improves an animal’s fitness, and that such behaviour tends to appropriately manage
internal resource levels under ‘natural’ ecological conditions.