Dialect Anthropol (2014) 38:333–351 DOI 10.1007/s10624-014-9355-0
A module is a module is a module: evolution of modularity in Evolutionary Psychology Giordana Grossi
Published online: 27 August 2014 Springer Science+Business Media Dordrecht 2014
Abstract The concept of modularity has been central in behavioral and neural sciences since the publication of Fodor’s The Modularity of Mind (1983). Fodor strived to explain the functional architecture of the mind based on the distinction between modular and central systems. Modular systems were deemed to have certain architectural features, such as automaticity, encapsulation, and domain specificity. Evolutionary psychologists have adopted the concept to characterize purportedly evolved human adaptations. In an influential paper, Barrett and Kurzban (Psychol Rev 113(3):628–647, 2006) proposed a definition of modules purely in terms of functional specialization. It is here argued that such strategy marks a shift in Evolutionary Psychology’s theoretical emphasis, as it trivializes the investigation of proximate causes in evolutionary theorizing; furthermore, it leaves the door open to too much flexibility on what counts as evidence for purportedly evolved modules. Keywords brain
Evolutionary Psychology Modularity Human psychology Human
History and definition of modularity Modularity is a central concept in the cognitive and neural sciences. Generally speaking, modularity is the idea that the mind and the brain are organized in systems that are specialized to process specific types of information and that are functionally independent (Coltheart 1999). The concept became predominant in psychology after the publication of Fodor’s The Modularity of Mind (1983). Fodor’s (1983) hypothesis about the functional architecture of the mind was proposed to explain the G. Grossi (&) Department of Psychology, State University of New York at New Paltz, 600 Hawk Drive, New Paltz, NY 12561, USA e-mail:
[email protected]
123
334
G. Grossi
types of psychological mechanisms underlying the different facts of mental life (e.g., sensation, perception, volition, and learning). In his account, the architecture of the mind is divided into modular and central systems. Modular systems include perceptual (or input) and linguistic mechanisms, whereas central systems pertain to decision-making systems, knowledge, and beliefs. The latter systems were deemed not to be modular and cut across a variety of cognitive domains. Without clearly defining the term ‘‘module,’’ Fodor associated modular systems with specific characteristics, such as automaticity, informational encapsulation, and domain specificity: •
•
•
Automaticity A system was conceptualized as automatic if its mechanisms are to be ‘‘obligatorily applied’’ (p. 53), such as those involved in spoken word recognition in native speakers of a language: Even when participants are asked to focus on low-level (acoustic–phonetic) characteristics of speech, it is impossible to turn off word recognition mechanisms. Informational encapsulation A system was deemed to be encapsulated if it is less permeable to background information, contextual factors, or beliefs than other systems. Fodor illustrated this property by discussing the mechanisms responsible for visual illusions. In the Mu¨ller–Lyer illusion, the line flanked by two arrow tails looks longer than the line flanked by two arrowheads. This perception persists even after the viewer is informed that the two lines are identical in length. Domain specificity Modules are specialized, in the sense that only a restricted class of stimuli has access to them. For example, the input systems involved in recognizing speech are different from the auditory systems involved in the recognition of non-speech sounds (Fodor 1983).
Modular systems tend to have other related characteristics, such as being hardwired (i.e., ‘‘associated with specific, localized, and elaborately structured neural systems,’’ p. 37) and innate (i.e., they ‘‘develop according to specific, endogenously determined patterns under the impact of environmental releasers,’’ p. 100). According to Fodor, modules needed not share all these characteristics; furthermore, modularity is a matter of degree. Fodor’s speculative view of the mind, born within philosophy, prompted a considerable amount of research and discussion among cognitive psychologists and neuroscientists. Much of the empirical research focused on determining how perceptual and linguistic systems function and whether they exhibit characteristics such as encapsulation, automaticity, and domain specificity. Empirical findings revealed a complex picture in which not even perceptual systems turned out to exhibit some of the defining characteristics of modularity. For example, some charged that input systems do not exhibit strict encapsulation because their functioning is modulated by attention (e.g., Wojciulik et al. 1998) or higher-level cognitive processes (Farah 1994). Furthermore, behavioral studies had already shown that perceptual mechanisms are modulated by contextual factors and information from other sensory modalities. For example, in the McGurk effect
123
A module is a module
335
(McGurk and MacDonald 1976), the auditory perception of a syllable (e.g., /ba/) changes when the syllable is visually presented with the face of a person pronouncing another sound (e.g., /ga/): In this case, participants hear /da/. These and other empirical results led to a more refined understanding of how the brain is organized and functions, as I will discuss later. Evolutionary Psychology1 and modularity In the late 1980s, the concept of modularity was borrowed by a group of evolutionary psychologists (e.g., John Tooby, Leda Cosmides, Steven Pinker, David Buss) and reframed in terms of evolved adaptations. In their view, the human mind is comprised of numerous specialized modules (or ‘‘mini-computers’’) that have evolved under selective pressure to solve recurrent problems in humans’ ancestral past (Cosmides and Tooby 1997, p. 11). Therefore, the Evolutionary Psychology (EP henceforth) version of modules rests on the assumptions that they are supported by dedicated neural circuits, shaped by natural selection, and genetically specified (Pinker 1997, p. 21). Examples of modules include face recognition, tool-use, fear, social-exchange, kin-oriented motivation, child-care, social-inference, sexualattraction, friendship, grammar acquisition, and theory of mind (Tooby and Cosmides 1992, p. 113). A number of assumptions are made regarding these modules. For example, they are domain-specific, that is, ‘‘relatively well-engineered for solving ancestral adaptive problems’’ (Cosmides and Tooby 1994, p. 88). This assumption relies on the claim that such problems cannot be solved by domain-general mechanisms. Evolutionary psychologists argue that different problems require sensitivity to different types of information and different solutions. For example, selecting food and selecting a mate typically lead to different solutions; therefore, they rely on different, and specialized, mechanisms (Confer et al. 2010). Another assumption is that the number of specialized modules has increased in humans, compared to other primates and mammals, and this increase ensures humans’ flexible and complex behavior. In this context, modules are ‘‘prespecified’’ structures, as they contain information about the world: The more a system initially ‘‘knows’’ about the world and its persistent characteristics, and the more evolutionarily proven ‘‘skills’’ it starts out with, the more it can learn, the more problems it can solve, the more it can accomplish. (Tooby and Cosmides 1992, p. 113) Evolutionary psychologists’ focus is on a computational, or functional, analysis of behavior, based on Marr’s distinction between levels of description (Cosmides and Tooby 2003; Cosmides and Tooby 1994). In his influential book Vision (1982/ 2010), Marr posited that informational systems can be analyzed at three different 1
The term ‘‘Evolutionary Psychology’’ will be used in this paper to denote the thought of an influential group of researchers including, among others, Leda Cosmides, John Tooby, David Buss, and Steven Pinker. It is not to be confused with ‘‘evolutionary psychology’’ as the general field of inquiry. Please see the Introduction to this issue for more information.
123
336
G. Grossi
levels of description: computational or functional (what a system does and why; the problem that the system solves); algorithmic/representational (how a system does what it does, what representations and processes it uses); and implementational (how a system is neurally/physically implemented). Marr considered the three levels of description ‘‘rather loosely related’’ (2010, p. 25) and constrained by different issues (independence among levels).2 For this reason, for some phenomena, there might be no unitary theory connecting all three levels. By privileging a computational analysis of behavior, evolutionary psychologists adopt reverse engineering to identify traits that have been advantageous (in terms of reproductive fitness) during human evolution in ‘‘ancestral times.’’ These ‘‘design features’’ are identified based on hypotheses concerning the problems faced by our ancestors and speculations about the ancestral environments in which they lived. Design features, which characterize evolved adaptations, can be recognized by three characteristics: They are specialized to solve an adaptive problem; their phenotypic characteristics are unlikely to have arisen by chance; and they are not explained as the byproduct of other mechanisms to solve other ancestral problems (Cosmides and Tooby 2003). It is important to note that many of the EP assumptions and claims (e.g., domain specificity or modularity of a given system) are rarely accompanied by supporting empirical data (a point already observed by others; e.g., Richardson 2007). For example, in Tooby and Cosmides’ writings, the existence of specific modules is typically assumed, not demonstrated. The existence of a behavior that is hypothesized to be evolutionary relevant is simply explained by the existence of a dedicated module: Humans’ use of tools is explained by the existence of a module specific for using tools; humans’ use of language characterized by grammar is explained by the existence of a grammar acquisition module. However, there is no convincing evidence for the existence, for example, of a grammar acquisition module. Children exhibit a predisposition to learn language, but this does not imply that they have a mechanism or a circuit specialized to learn grammar per se. Indeed, researchers have shown that, in children, grammatical complexity is tied to vocabulary size, indicating that the acquisition of grammar relies on other aspects of linguistic communication (Bates and Goodman 1997). In addition, children whose language reveals grammatical difficulties also exhibit a range of other linguistic and/or non-linguistic problems (e.g., Karmiloff-Smith 2006; Vargha-Khadem et al. 1995). Therefore, the claim that humans have a grammar acquisition module remains a presupposition. Furthermore, the characteristics of such hypothesized modules remain vague and underspecified. For example, speaking of depth perception, Tooby and Cosmides (1992) remarked that, ‘‘Our depth perception mechanism has this property, for example: It works well because it combines the output of many small modules, each sensitive to a different cue correlated with depth.’’ (p. 104). These descriptions, typically shared without pertinent details from the empirical literature, inevitably 2
Marr’s distinction among levels of description and the importance he gave to questions of function were not informed by evolutionary theory but by work on brain-lesioned patients in human neuropsychology. For a discussion, see Shapiro and Epstein (1998).
123
A module is a module
337
raise issues of circularity, as what needs to be demonstrated is actually included in the premises (depth perception works well because the system picks up cues of depth). Numerous scholars have raised issues regarding the concept of modularity adopted by evolutionary psychologists. For example, focusing on cognitive architecture, Shapiro and Epstein (1998) have remarked that whether or not a system is domain-specific or domain-general remains a matter of empirical evidence, not theoretical assumptions. Based on empirical data, DeSteno et al. (2002) have questioned the evolutionary meaning of the sex differences in jealousy reported by Buss et al. (1992) and the existence of the ‘‘jealousy module.’’ Other researchers have rejected mass modularity (the assumption that the mind is mostly comprised by domain-specific modules) on neuroscientific grounds (Bolhuis et al. 2011; Buller 2005). Such critiques have raised counter-arguments from some evolutionary psychologists (e.g., Barrett et al. 2006). In an influential paper, Barrett and Kurzban (2006; B&K henceforth) claimed that these critiques often misrepresent the concept of modularity proposed by evolutionary psychologists. The authors emphasize that that EP version of modularity centers on the notion of ‘‘functional specialization’’: Here we argue and provide evidence for the view that constructive progress has been undermined by the fact that opponents of modern views of modularity have critiqued modern positions as though the original (Fodorian) conception of modularity were intended. We also assert, as have other evolutionary psychologists (Barrett 2005; Cosmides and Tooby 1994; Pinker 1997; Sperber 1994; Tooby and Cosmides 1992, 2005), that a broader notion of modularity than the one Fodor advanced is possible: in particular, a modularity concept based on the notion of functional specialization, rather than Fodorian criteria such as automaticity and encapsulation (pp. 628–629, italics in the original). This position has been echoed by others, such as Confer et al. (2010), who claimed that ‘‘psychological adaptations are not separate ‘modules’ in the Fodorian (Fodor 1983) sense of informational encapsulation; rather, they often share components and interact with each other to produce adaptive behavior…’’3 (p. 111; see also Cosmides and Tooby 2003, p. 63, for a similar position). After showing that, despite B&K’s claim, evolutionary psychologists have described modules or specific information-processing systems in terms of Fodorian characteristics,4 I will argue that the strategy of defining modules only in terms of functional specialization is problematic. Indeed, by stripping them of defining architectural features or treating them loosely and vaguely, such strategy reduces constraints on hypothesis testing and causes too much flexibility when researchers test hypotheses regarding the modularity of specific systems. Here, I will discuss 3
It is important to note that such interactions are also assumed in EP: When and how they take place is typically not a matter of empirical exploration but post hoc speculation (see Lickliter and Honeycutt 2003, for a similar point).
4
For a similar argument, see Chiappe and Gardner (2012).
123
338
G. Grossi
these issues by focusing on two modular features: domain specificity and neural specificity.
Fodorian modularity in Evolutionary Psychology Contrary to B&K’s claims, work in EP has in fact focused on Fodorian characteristics of modularity. In most cases, the reference to Fodor’s features is not explicitly acknowledged, but the research strategy adopted by the authors clearly indicates a reliance on them. For example, the concept of automaticity surfaces in Buss et al.’s (1992) study on sex differences concerning jealousy, as they measured changes in autonomic nervous system responses triggered by infidelity scenarios. Jealousy is ‘‘triggered’’ in a variety of papers (Buss 1998; Buss and Haselton 2005; Buss et al. 1996, 1999); furthermore, male jealousy is considered to be ‘‘obligate’’ in long-term relationships (Buss 1998, p. 26). The ‘‘fear module’’ ¨ hman (2002) is ‘‘selective, automatic, and hypothesized by Mineka and O encapsulated’’ (that is, ‘‘preferentially activated by stimuli related to survival threats in evolutionary history…. automatically activated by fear-relevant stimuli … relatively impenetrable to conscious cognitive control…’’, p. 927). Similarly, New et al. (2007) have hypothesized the existence of a category-specific attention mechanism aimed at automatically checking the status of people and animals in environment. They used a well-known paradigm (change blindness) to test whether changes concerning humans or animals in a scene are found more easily than changes concerning non-animate objects. The evidence of specialized neural mechanisms has also been taken as a marker of modularity by evolutionary psychologists. Duchaine et al. (2001) provided a review of ‘‘some of the recent evidence for functionally specialized problem-solving machinery in the brain’’ (p. 225). Tooby and Cosmides (2005) remark that specialized mechanisms rely on ‘‘neurocognitive specializations’’ that are ‘‘functionally and neurally distinct from more general abilities to process information or behave intelligently’’ (p. 587) and have attempted to find evidence of a cognitive mechanism specialized to identify cheaters in social exchanges (cheater-detection module) at the neural level (Stone et al. 2002). Jung et al. (2012) claim that perception of facial attractiveness is a ‘‘premier’’ example of psychological adaptation and cite, among other forms of evidence, the fact that it is associated with the activation of specialized brain areas. Confer et al. (2010) mention the ‘‘sexdifferentiated patterns of neural activation’’ (p. 114) associated with jealousy (found by Takahashi et al. 2006) as a support for the tenet that sex differences in jealousy have evolved. These are just some examples that demonstrate that B&K’s claim that evolutionary psychologists have not relied on Fodorian definitions of modularity is incorrect. Therefore, it is not clear how ‘‘… constructive progress has been undermined by the fact that opponents of modern views of modularity have critiqued modern positions as though the original (Fodorian) conception of modularity were intended’’ (p. 628). If anything, constructive progress has actually been made thanks to studies that have tested the modularity of purported evolved
123
A module is a module
339
modules and their development. These studies have complicated the picture provided by evolutionary psychologists. For example, according to Confer et al. (2010), the existence of ‘‘evolved fear adaptations,’’ such as fear of snakes, is relatively uncontroversial, as shown by evidence in terms of automaticity, neural specificity, and difficulty of extinction. Also, fear of snakes exists in both humans and other primates, whereas humans do not exhibit fear of objects that are historically more recent, even if more dangerous than snakes (e.g., guns, electric outlets): A programmatic series of studies has shown that an intense fear of snakes exists in humans and other primates; snakes and spiders embedded in complex visual arrays automatically capture attention far more than do harmless objects—they ‘‘pop out’’ of the visual array; humans rapidly condition to fear snakes more than most other stimuli; and the snake fear adaptation is selectively and automatically activated, it is difficult to extinguish, and it can be traced to specialized neural circuitry… (p. 111). However, research has shown that objects like guns capture as much attention as snakes in spatial attention experiments (Carlson et al. 2009) and visual search tasks in which participants are asked to find a target object in an array of distractors (e.g., Blanchette 2006; Brosch and Sharma 2005; Fox et al. 2007). Tipples et al. (2002) also showed that non-threatening animals such as kittens and bunnies are processed similarly to snakes in visual search experiments. Therefore, conclusions from studies in which snakes were presented among non-animal pictures were drawn incorrectly, as a threat manipulation was confounded with animateness. Furthermore, Hugdahl and Johnsen (1989) showed that people can be rapidly conditioned to guns pointed at them and this conditioned response is as difficult to extinguish as the response to snakes. As for the existence of ‘‘specialized neural ¨ hman (2002) (cited by Confer et al. 2010) discuss the role of circuits,’’ Mineka and O the amygdala in fear conditioning, not specific neural circuits for fear of snakes (which have never been identified, to my knowledge). Finally, fear of snakes is not present in all humans and is acquired in monkeys based on specific and non-obvious experiences, as developmental studies have shown. Mineka et al. (1984) had already noted that laboratory-reared monkeys do not exhibit fear of snakes, but will exhibit it after observing their wild-rear parents’ reactions to snakes. In a remarkable and rarely cited study, Masataka (1993) showed that fear of snakes was present in laboratory-reared monkeys fed with fruit and live insects but absent in laboratoryreared monkeys fed on fruit and monkey chow. It appears that fear of snakes is acquired in monkeys based on a variety of experiences, for example observing their parents or encountering and feeding on live insects (part of the normal monkey’s diet in the wild). Monkeys do not develop a fear of snakes without these experiences. Similarly, Berger et al. (2001) have shown that moose’s fear of wolves and bears (moose’s ‘‘natural predators,’’ based on their evolutionary history) disappeared in animals where these predators were decimated within a few generations (1959–1992). When predators were reintroduced, fear was relearned in one generation through specific experiences, such as the loss of offspring, and then transmitted culturally to the next generation. In summary, this body of research
123
340
G. Grossi
sheds light on the actual experiences and mechanisms that, within each generation, foster the development of fear of snakes or predators. Explaining such behavior simply in terms of an evolved adaptation or module does not capture its complexity. Confer et al. (2010) use the term ‘‘evolved adaptions’’ to refer to fear of snakes and other evolutionary-relevant faculties (e.g., jealousy). They maintain that such a concept does not coincide with Fodor’s concept of modularity (p. 111). However, the characteristics that they list to support the existence of a snake fear adaptation mostly overlap with the characteristics proposed by Fodor (1983). Therefore, such features are used as evidence for modularity in some cases and dismissed in others, in particular when experiments do not provide support for them (e.g., Barrett et al. 2006). In their paper, B&K propose to remove these features from the definition of modularity or to use them flexibly, as discussed next.
From Fodorian modularity to untethered functionalism Domain specificity B&K’s view of modularity stems from Pinker (1997), ‘‘who argued that modules should be defined by the specific operations they perform on the information they receive, rather than by a list of necessary and sufficient features …’’ (p. 629). Modules are ‘‘… functionally specialized mechanisms with formally definable informational inputs’’ (p. 630). They have design features that evolved by natural selection ‘‘acting on the developmental systems that build modules during development’’ (p. 630). Through natural selection, they become domain-specific, as they process specific information in a specific way (they have ‘‘specific input criteria,’’ p. 630). For example, ‘‘systems specialized for assessing the numerosity of objects accept only representations previously parsed into distinct objects; systems specialized for speech perception process only transduced representations of sound waves; and systems specialized for making good food choices process only representations relevant to the nutritional value of different potential food items’’ (p. 630). B&K’s concept of domain specificity seems not to differ substantially from Fodor’s. However, domain is here conceptualized in terms of ‘‘formal properties of representations’’: …, we wish to stress that we intend the broadest construal of the term domain to include, in principle, any possible means of individuating inputs. We do not intend a reading of domain as content domain, in the folk sense of domains individuated by the meaning of their constituents. Rather, we define domains as individuated by the formal properties of representations because, we believe, this is the only possible means by which brain systems could select inputs. As a corollary, by virtue of the fact that formal properties determine which inputs are processed, a mechanism specialized for processing information of a particular sort can, as a by-product, come to process information for which it was not originally designed,… (p. 630)
123
A module is a module
341
These claims raise some issues. First of all, how are formal and non-formal properties defined, empirically identified, and distinguished from each other? These questions assume relevance when we try to identify the systems specialized for a given function. For example, what are the systems specialized for speech perception, and what are their formal properties? Importantly, how do they differ from the formal properties of non-speech perceptual auditory systems? To state, as B&K do, that ‘‘systems specialized for speech perception process only transduced representations of sound waves’’ (p. 630) is to state the obvious, as speech is a signal based on sound waves, and transduction is the translation of a form of energy (mechanical, in auditory perception) into neural signals. The key word in their claim is, perhaps, ‘‘only,’’ but without clarifying what the specific systems are and how they work, the claim of input specificity remains vague and unsubstantiated.5 A second, and related, issue concerns the input criteria. Systems can accept only specific types of input, but they can process information for which they were not originally designed. According to B&K, the fact that a system is activated by other types of input (that is, lacks domain specificity) does not undermine its modularity. Specificity of input is then substituted with ‘‘vocabulary on inputs’’: …, if every cognitive mechanism has specifiable information that it accepts as inputs, even if some systems accept information in multiple formats (or can be controlled or influenced horizontally or top-down by other systems), then the crucial issue is the vocabulary of inputs a given mechanism accepts. No mechanism is either encapsulated or unencapsulated in an absolute sense. Cognitive mechanisms can be referred to as encapsulated with respect to certain information types but not others. What is important is to specify how information is accessed and how it is processed, including the input criteria that must be met for processing to occur. (p. 631, italic added) This move is seen as necessary: The focus on input conditions and function clarifies what is meant by ‘‘domains,’’ because a theory of function will constrain hypothesized formal input conditions for information-processing devices. Sperber (1994) referred to the ‘‘proper’’ domain of a module as the class of inputs the module was designed by natural selection to process. For example, the proper domain of a face recognition system would be, putatively, faces of conspecifics (Kanwisher 2000). The ‘‘actual domain’’ might be, and indeed in many cases must be, a broader class of tokens than the type for which certain modular systems evolved: for example, perhaps not only faces but the wider set of stimuli that 5
As previously discussed, speech perception is indeed a complicated affair. It is affected by the presentation of both visual (McGurk and MacDonald 1976) and tactile (Gick and Derrick 2009) stimuli. Such influence is highly specific and was hypothesized to occur based on an in-depth knowledge of how speech perception systems work. For example, inaudible air puffs applied on participants’ right hand or neck while they listened to syllables caused participants to mishear ‘b’ (non-aspirated) as ‘p’ (aspirated; Gick and Derrick 2009). These results demonstrate that perceivers integrate event-relevant tactile information in auditory perception and provide evidence for an extensive degree of integration across sensory modalities.
123
342
G. Grossi
have formal properties that cause them to be processed by the face recognition system. In addition, a view of evolved function informs hypotheses about inputs, including the contextual mediation of processing, as information about context can itself be an input to modular systems …. (p. 631) B&K maintain that specialization is identified by the presence of specific input criteria, which are identified by an analysis of function. It is the input that indicates what types of information is processed by a system. However, this input might include other aspects of stimuli that share ‘‘formal properties’’ with the specialized input. The authors do not address the conditions under which one stimulus or a broader class of tokens characterizes the ‘‘actual domain’’ of a module. Why ‘‘must’’ we have, in some cases, a variety of inputs and in others a specific type of input? This lack of specificity opens the door to too much flexibility on what counts as evidence for purportedly evolved modules. With respect to faces, B&K do not discuss what the formal properties shared by other visual objects might be. Evolutionary psychologists’ hypotheses are typically driven by hypothesized evolutionary scenarios; as a consequence, they do not focus on how the information is actually accessed or processed. Specific hypotheses are rarely made. For example, it is difficult to imagine how a functional analysis of face perception or the hypothesis that the face recognition system has evolved to allow humans to recognize their conspecifics might explain and predict some recent empirical findings. Specifically, in car experts, face processing is disrupted by the visual presentation of cars (Gauthier et al. 2003; see also McKeeff et al. 2010). Furthermore, the neural region of the temporo-occipital cortex associated with face processing (face fusiform area, Kanwisher and Yovel 2006) is significantly activated by objects of expertise: The presentation of cars and birds activates this area in car and bird experts, respectively (Gauthier et al. 2000); furthermore, this area is activated in individuals who became expert in recognizing exemplars of a new made-up category (i.e., Greebles; Gauthier et al. 1999). The question of what purported formal properties are shared by faces and cars in car experts, or by faces and birds in bird experts, must be addressed to explain these findings. In cognitive neuroscience, these data have been used to challenge the domain specificity hypothesis proposed by Kanwisher, according to whom face perception is carried out by neural systems specialized for processing faces (e.g., Kanwisher and Yovel 2006). Gauthier and colleagues proposed that faces might preferentially activate specific brain regions not because they are specialized for faces per se, but because they involve a level of visual representation that is not typical of other object categories (i.e., processing information beyond single parts, or holistic processing, and individuation of exemplars, not just types; Gauthier and Tarr 2002). Therefore, functional specialization is driven by expertise, not faces per se. B&K might propose that both theories can be incorporated within an evolutionary framework where objects of expertise share some formal properties with faces (for an attempt to integrate them, see Barrett 2012). However, without a definition of ‘‘formal properties’’ (based on physical characteristics? type of processing?) and an in-depth knowledge of how a given system works, this a posteriori integration would remain problematic. The underspecification of a system’s characteristics
123
A module is a module
343
(here in terms of input and processes) makes the theory amenable to subsequent adjustments by following a postdiction logic. In order for theories to be useful, they should explicitly identify what type of empirical findings would falsify them. If any data can be adjusted to fit a theory, the theory becomes too powerful and ultimately superfluous. Therefore, the concern is that the reformulation of the concept of modularity in terms of functional specialization leaves too much leeway on the type of evidence needed to support the hypothesis that a given system is modular. It does not matter what features a system has; as long as it has design characteristics that one might define as functional, based on hypothesized selective environmental pressures, the system is considered a mechanism evolved through natural selection. The theory itself then becomes sufficient reason (or evidence) to establish whether or not a mechanism is considered to be specialized and adaptive. This is a critical issue as studies devised to test EP theories typically do not test the hypothesis that a trait was adaptive in ancestral times, but establish the presence of some sort of effect, ability, or group difference that is consistent with the hypothesis (e.g., Bolhuis et al. 2011; Richardson 2007). The effect is in turn taken as evidence for the hypothesis that an adaptation has been found (see also Lloyd 1999, and Nash, this issue). However, an effect or a group difference says nothing about their origins. Neural specificity B&K apply the same flexible logic to another important criterion for modularity, that is, specificity of neural circuits (Fodor’s ‘‘hardwiring’’). Specificity of input entails specificity at the neural level. If it is true that a module needs not to be localized in a single and small area of the brain (B&K, p. 641), its neural implementation should be clearly defined in terms of brain regions involved, input projections and output. While some EPs have researched the neurophysiological underpinning of purported modules (e.g., Stone et al. 2002), examination of brain mechanisms to substantiate hypothesis regarding evolved modules has been rare in EP (see Bolhuis et al. 2011, for a similar point). Specialization of circuits is typically assumed. On the one hand, B&K concede that isomorphism between functional and structural architecture exists; on the other hand, ‘‘at a larger, macroscopic level, there is no reason to assume that there must be spatial units or chunks of brain tissue that neatly correspond to information-processing units.’’ (p. 641). As is typical in EP papers, instead of providing scientific evidence for such claim, the authors present an analogy and discuss how, in an electronic device, Individual wires have specific functions, but at the level of the entire machine, wires with different functions might cross and overlap. For this reason, removing, say, a three-inch square chunk from the machine would not necessarily remove just one of the machine’s functions and leave the rest intact. In brain terms, it could be, and probably is, that macroscopic regions of brain tissue include neurons from multiple information-processing systems with multiple functions… (p. 641)
123
344
G. Grossi
Based on such analogy, B&K are ready to dismiss evidence from neuroimaging and brain-lesion studies: Here, we argue that modularity in the sense of functionally specialized information processing can exist even in the absence of evidence of spatial localization from, for example, fMRI or lesion studies.6 (p. 641) While it is true that large lesions typically compromise a variety of functions, that modules need not be localized in one single region, and that neuroimaging techniques have limitations in terms of spatial resolution (besides, they provide an indirect measure of brain activity, as they measure changes in blood flow metabolism; Friston 1997), B&K fail to provide alternative methods to investigate how modular systems are neurally implemented. Their strategy consists in dismissing the usefulness of two major methods used by cognitive neuroscientists to test models of neural architecture and implicitly suggesting that such an enterprise might be worthless when claims of modularity are tested within an EP framework. Therefore, we are asked to evaluate hypotheses about evolved modular systems, characterized by specific neurally circuits, without a scientific investigation of their neurophysiological underpinnings or correlates. Similarly, activation of the same neural system by two different sets of stimuli is rejected by B&K as evidence for shared mechanisms, based on the limitation of neuroimaging techniques in terms of spatial resolution: the two stimuli might activate different populations of neurons even in a very circumscribed region7 (p. 641). However, there are other techniques and research paradigms that can be used to overcome the localization issues mentioned by B&K. For example, Gauthier et al. (2003) employed a different methodology to investigate whether neural regions for faces and objects of expertise are functionally independent. They adopted an interference paradigm where participants were asked to process two stimuli at the same time. If processing two types of stimuli relies on shared cognitive and neural resources, performance is negatively impacted (for example in terms of response time) compared to a condition in which the two types of stimuli draw from different resources. Gauthier et al. (2003) reasoned that face processing would be disrupted by the presentation of cars in car experts if faces and cars were processed by the same brain circuits. This is exactly what the authors found. Importantly, the interference effect arose during the early stages of visual processing. The authors recorded event-related potentials, or ERPs, while participants performed the experiment. ERPs are voltage fluctuations in the electroencephalogram elicited by the presentation of a controlled stimulus. The latencies of different positive and 6
Contrast this position with that of Price and Friston (2005), who call for a rigorous investigation before claiming that a brain region exhibits functionally specialization: ‘‘To infer functional specificity requires a demonstration that an area is activated only by tasks that engage its function and no others’’ (p. 265).
7
B&K consider informative the activation of different brain areas by different types of stimuli to establish that the two stimulus sets engage different systems, and I agree. However, even in this case, careful investigation is necessary to establish the meaning of differential brain activation, as it might reflect a variety of factors (e.g., differences in non-critical physical characteristics, levels of practice, levels of attention). In order to identify what brain regions are necessary for a given function, functional specialization at the neural level needs to be investigated.
123
A module is a module
345
negative components in an ERP reveal the time course of activation (in the order of milliseconds) of the neuronal populations that are recruited during the processing of that stimulus (Picton et al. 2000). Asking car experts to simultaneously process faces and cars influenced the amplitude of a component associated with the early stages of face perception, known as the N170, elicited over posterior regions between 140 and 190 ms. Therefore, in car experts, processing of cars and faces relies on shared neural circuits or, in the authors’ words, ‘‘are not functionally independent.’’ (p. 431). These results cast doubts on the hypothesis that these brain regions are functionally specialized for faces per se. However, B&K still maintain that such systems are functionally specialized for faces: Unlike intuitive ideas about domain specificity, functional specificity affords a means of individuating functional domains using evolutionary functionalist principles to do so (in this case, for example, the importance of faces as a domain is suggested by the fitness benefits of being able to recognize individual conspecifics; Cosmides and Tooby 1994; Tooby and Cosmides 1992). (p. 634) However, the cited works do not provide independent evidence of the ‘‘fitness benefits’’ related to recognizing conspecifics. These benefits are only assumed to have existed during early human evolution, raising once again issues of circularity (see Nash, this issue).
Modularity, thirty years later Fodor (1983) proposed the modular framework as a working hypothesis to understand the functional organization of the mind. Such organization relied on the distinction between modular and non-modular systems. Modular systems were hypothesized to have specific characteristics and architectural constraints. This hypothesis prompted a considerable amount of empirical research aimed at testing the modularity of a variety of cognitive systems. These studies have refined our understanding of how the brain is organized and functions, as empirical findings revealed a complex picture in which not even perceptual and linguistic systems turned out to exhibit the defining characteristics of modularity (e.g., Farah 1994; McGurk and MacDonald 1976; Price et al. 2005; Wojciulik et al. 1998). The concept of informational encapsulation further appeared to break down at the neural level, as revealed by studies on multisensory integration during early perceptual processing (for a review, see Ghanzanfar and Schroeder 2006). Furthermore, functional neuroimaging data revealed no reliable one-to-one mapping between cognitive functions and specific neural regions (Price and Friston 2005). This empirical work led to a redefinition of the architecture of the mind and the brain, mostly in terms of interconnectivity, where specialization of a region is not an intrinsic property but depends on connections with other regions (e.g., Friston and Price 2001; McIntosh 2000).
123
346
G. Grossi
At the same time, work in developmental cognitive neuroscience has shown that the brain organization found in adults emerges during development (e.g., Cohen Kadosh and Johnson 2007; Neville and Mills 1997). This development is contextdepended and shaped by a variety of factors at different levels of organization: from the activity within and between single neurons, to the activity of different brain regions, to the behavior of the organism itself, in turn influenced by its environment (Westermann et al. 2007). Changes at one level can dramatically alter brain organization, as shown by the reorganization of the cortex documented in individuals with congenital blindness (e.g., Amedi et al. 2005), whose visual areas are recruited to support linguistic functions during speech production. Therefore, an approach where levels of descriptions are seen as independent (Marr 1982/2000) is untenable as brain organization and cognitive development emerge through local and interactive processes. In this context, cognitive and brain systems that are specialized in adults develop in a highly interconnected brain where regions co-develop with other brain regions, not in isolation. What a brain region or neuron does, in terms of function, depends on its interaction with other regions and neurons (e.g., Friston and Price 2001; Price and Friston 2005); it even depends on the state of distributed neural networks (e.g., Fontanini and Katz 2008). Within this framework, the specialization of neural systems (modularity) assumes a different meaning, one that is anchored into the physical system of a developing organism (Karmiloff-Smith 2013). Development is an interactive phenomenon, or, in Balaban’s (2006) words, ‘‘a conversation, not a blueprint’’ (p. 310), where the characteristics of an organism, including systems considered modular, are constructed here and now. Current contextual influences are not factored into EP, except in terms of conjectures about recurrent selective pressures from the ancestral past. Therefore, the empirical research conducted after the publication of The Modularity of Mind has put the concept of modularity in perspective, contextualized it, and redefined it (in terms of partial and emergent specialization). At the same time, it has given it a more defined place, one that is rooted in the actual mechanisms that characterize neural and cognitive development. Such research focused on testing, not assuming, modularity. In contrast, the modules proposed by evolutionary psychologists have remained elusive, disembodied, unanchored to a rigorous analysis of proximate causes, including their neural underpinnings. B&K’s reframing of the concept in terms of functional specialization exacerbates this elusiveness. B&K do not consider it necessary to provide evidence of structural modularity to show that a system is an evolved module: the adoption of functional specialization as the defining feature of modularity accommodates a flexible form of domain specificity, which can be used as needed. Flexibility is further adopted in establishing evidence of neural specificity for evolved adaptations. Using Karmiloff-Smith’s (2006) words, they want to have their modular cake and eat it too (p. 565). Indeed, in a later paper, Barrett and Kurzban (2012) adopt a definition of modularity that strips purportedly evolved modules of any architectural constraints: ‘‘if X is a mechanism, and if it has a design (i.e., has been shaped by the process of natural selection acting over evolutionary time), then it is what we are calling a ‘module’’’ (pp. 684–685, italics in the original). This definition includes systems
123
A module is a module
347
that have traditionally been considered non-modular according to Fodorian characteristics. It is perhaps ironic that a concept so central to the EP approach has become so ineffable. Humans supposedly have many instincts, many more than other animals (Cosmides and Tooby 1997), and specific neurocognitive mechanisms to support them. Yet, evidence of modularity for such mechanisms is based less on reliable evidence of architectural features and more on functional guesses based on hypothesized evolutionary scenarios. When issues of modularity are raised (e.g., DeSteno et al. 2002), key constraints are not considered critical (e.g., Barrett et al. 2006) or removed from the definition of modularity (Barrett and Kurzban 2006, 2012). The adoption of words like machine and replicators (Tooby and Cosmides 2005), along with the recurrent (and ill-conceived) brain-as-computer metaphor, reveals a progressive estrangement from the study of biological organisms as real and embodied, and stands in stark contrast with the call for an integrated study of behavior by computational neuroscientists. Indeed, in his Afterword to the 2010 edition of Marr’s Vision, Poggio remarked how the connections between levels must be reemphasized: (1) insights gained on higher levels help us ask the right questions and do the right experiments at the lower levels, and (2) it is necessary to study nervous systems at all levels simultaneously. From this perspective, the importance of coupling experimental and theoretical work in the neurosciences follows directly: without close interaction with experiments, theory is very likely to be sterile (p. 364). The trend to isolate functional analyses from the investigation of proximate causes of behavior (such as its physical implementation and development)8 is evident in other recent writings. For example, in a paper addressing the relationship between exposure to disease threats and preference for physically attractive leaders, White et al. (2013) write that such relationship was predicted based on an evolutionary perspective and that, ‘‘… predictions couched in terms of more proximal mechanisms do not necessarily constitute ‘alternatives.’ Instead, it may be through these mechanisms that more distal processes have their effects’’ (p. 2435; see Barrett 2012, p. 10739, for a similar point). In this view, the study of the neural underpinnings of behavior has the purpose of simply revealing how distal functions are implemented; it does not inform the scientist, and it provides information that can only validate, not challenge, a functional analysis. Thus, it becomes irrelevant. As B&K have stated: ‘‘We remain agnostic with regard to the way that functional specificity is implemented in the brain and look forward to the accumulating evidence from developmental neuroscience to inform the details of these processes.’’ (p. 642). It is difficult to reconcile such a position with the claim that 8
I am not endorsing a rigid distinction between levels of explanation, as the utility of such distinction has been questioned in the light of new empirical findings (see Laland et al. 2011; Bateson and Laland 2013 for recent discussions). What I am arguing is that the study of proximal causes (development and neurobiological processes) cannot be neglected in evolutionary sciences, as such processes have been shown to shape evolutionary pathways (Lickliter and Honeycutt 2003).
123
348
G. Grossi
evolutionary psychologists identify ‘‘the proximate causal mechanisms of human behavior in the brain’’ (Barrett 2008, p. 173). B&K’s words seem to suggest a theoretical shift in EP, one that is associated with the abandonment of rigorous analyses at the implementational level and the retreat into a functionalist framework. This shift reveals a fundamental contradiction within EP. Some evolutionary psychologists have adopted concepts developed in other theoretical approaches (e.g., probabilistic epigenesis, Gottlieb 1998; progressive modularization, Karmiloff-Smith 2006; developmental system, Oyama 2000) and incorporated recent empirical findings from cognitive, developmental, and brain sciences in their theoretical contributions (e.g., Barrett 2012; Ellis and Bjorklund 2005). This incorporation demonstrates how important the analysis of proximate causes is for evolutionary theorizing. However, evolutionary psychologists do not implement such an integrative approach in their empirical investigations and rarely contribute to our understanding of evolutionary-relevant behaviors in terms of developmental and neural processes (see also Nash, this issue, and Bolhuis et al. 2011). B&K’s proposal exacerbates this trend. Until such integration is reached, rather than reframing modularity in terms of functional specialization, evolutionary psychologists should refrain from using the term when they refer to purportedly evolved adaptations. Barrett and Kurzban (2012) stated that they ‘‘would be happy to discard the term in favor of some mutually agreed upon term: ‘psychological specialization,’ ‘mental adaptation,’ ‘evolved cognitive mechanism,’9 or any equivalent that designates an aspect of mental structure that was shaped by natural selection’’ (p. 684). Following their suggestion, they should adopt a terminology that clearly acknowledges the focus on the functional (computational) level and removes any reference to architectural constraints and neural processes, as these are not investigated systematically. The vagueness and disconnect of EP theories will then emerge, along with the contradictions intrinsic to an approach that claims to treat psychology as a branch of biology ‘‘… that studies (1) brains, (2) how brains process information, and (3) how the brain’s information-processing programs generate behavior’’ (Cosmides and Tooby 1997, p. 5). Acknowledgments Many thanks to Alison Nash, Suzanne Kelly, Gowri Parameswaran, and two anonymous reviewers for their thoughtful comments and suggestions on an earlier version of this manuscript.
References Amedi, A., L. Merabet, F. Bermpohl, and A. Pascual-Leone. 2005. The occipital cortex in the blind: Lessons about plasticity and vision. Current Directions in Psychological Science 14(6): 306–311. Balaban, E. 2006. Cognitive developmental biology: History, process and fortune’s wheel. Cognition 101: 298–332. Barrett, H.C. 2005. Enzymatic computation and cognitive modularity. Mind and Language 20: 259–287. Barrett, H.C. 2008. Evolved cognitive mechanisms and human behavior. In Foundations of evolutionary psychology, ed. C. Crawford, and D. Krebs, 173–189. New York, NY: Lawrence Erlbaum Associates. 9
See Gantt et al. (2012) for a critique of the use of the term ‘‘mechanism’’ in EP.
123
A module is a module
349
Barrett, H.C. 2012. A hierarchical model of the evolution of human brain specializations. Proceedings of the National Academy of Science 109: 10733–10740. Barrett, H.C., D. Frederick, M. Haselton, and R. Kurzban. 2006. Can manipulations of cognitive load be used to test evolutionary hypotheses? Journal of Personality and Social Psychology 91(3): 513–518. Barrett, H.C., and R. Kurzban. 2006. Modularity in cognition: Framing the debate. Psychological Review 113(3): 628–647. Barrett, H.C., and R. Kurzban. 2012. What are the functions of System 2 modules? A reply to Chiappe and Gardner. Theory & Psychology 22(5): 683–688. Bates, E., and J.C. Goodman. 1997. On the inseparability of grammar and the lexicon: Evidence from acquisition, aphasia, and real-time processing. Language and Cognitive Processes 12(5/6): 507–584. Bateson, P., and K.N. Laland. 2013. Tinbergen’s four questions: An appreciation and an update. Trends in Ecology & Evolution 28(12): 712–718. Berger, J., J.E. Swenson, and I.-L. Persson. 2001. Recolonizing carnivores and naı¨ve prey: Conservation lessons from Pleistocene extinctions. Science 291: 1036–1039. Blanchette, I. 2006. Snakes, spiders, guns, and syringes: How specific are evolutionary constraints on the detection of threatening stimuli? The Quarterly Journal of Experimental Psychology 59(8): 1484–1504. Bolhuis, J., G. Brown, R. Richardson, and K.N. Laland. 2011. Darwin in mind: New opportunities for evolutionary psychology. PLoS Biology 9(7): 1–8. doi:10.1371/journal.pbio.1001109. Brosch, T., and D. Sharma. 2005. The role of fear-relevant stimuli in visual search: A comparison of phylogenetic and ontogenetic stimuli. Emotion 5(3): 360–364. Buller, D.J. 2005. Adapting minds: Evolutionary psychology and the persistent quest for human nature. Cambridge, MA: MIT Press/Bradford Books. Buss, D.M. 1998. Sexual strategies theory: Historical origins and current status. The Journal of Sex Research 35(1): 19–31. Buss, D.M., and M.G. Haselton. 2005. The evolution of jealousy. Trends in Cognitive Science 9(11): 506–507. Buss, D.M., R.J. Larsen, D. Westen, and J. Semmelroth. 1992. Sex differences in jealousy: Evolution, physiology, and psychology. Psychological Science 3: 251–255. Buss, D.M., R.J. Larsen, and D. Westen. 1996. Sex differences in jealousy: Not gone, not forgotten, and not explained by alternative hypotheses. Psychological Science 7(6): 373–375. Buss, D.M., T.K. Shackelford, L.A. Kirkpatrick, J. Chloe, H.K. Lim, M. Hasegawa, T. Hasegawa, and K. Bennett. 1999. Jealousy and beliefs about infidelity: Tests of competing hypotheses in the United States, Korea, and Japan. Personal Relationships 6: 125–150. Carlson, J.M., A.L. Fee, and K.S. Reinke. 2009. Backward masked snakes and guns modulate spatial attention. Evolutionary Psychology 7(4): 534–544. Chiappe, D., and R. Gardner. 2012. The modularity debate in evolutionary psychology. Theory & Psychology 22(5): 669–682. Cohen Kadosh, K., and M.H. Johnson. 2007. Developing a cortex specialized for face perception. Trends in Cognitive Sciences 11(9): 367–369. Coltheart, M. 1999. Modularity and cognition. Trends in Cognitive Sciences 3(3): 115–120. Confer, J.C., J.A. Easton, D.S. Fleischman, C.D. Goetz, D.M. Lewis, C. Perilloux, and D.M. Buss. 2010. Evolutionary psychology: Controversies, questions, prospects, and limitations. American Psychologist 65(2): 110–126. Cosmides, L., and J. Tooby. 1994. Origins of domain-specificity: The evolution of functional organization. In Mapping the mind: Domain specificity in cognition and culture, ed. L. Hirschfeld, and S. Gelman, 85–116. New York: Cambridge University Press. Cosmides, L., and J. Tooby. 1997. Evolutionary psychology: A primer. Retrieved January 7, 2014, from http://www.cep.ucsb.edu/primer.html. Cosmides, L., and J. Tooby 2003. Evolutionary psychology: Theoretical foundations. In Encyclopedia of cognitive science, ed. L. Nadel, 54–64. London: Macmillan. DeSteno, D., M.Y. Bartlett, J. Braveman, and P. Salovey. 2002. Sex differences in jealousy: Evolutionary mechanism or artifact of measurement? Journal of Personality and Social Psychology 83(5): 1103–1116. Duchaine, B., L. Cosmides, and J. Tooby. 2001. Evolutionary psychology and the brain. Current Opinion in Neurobiology 11(2): 225–230. Ellis, B.J., and D.F. Bjorklund (eds.). 2005. Origins of the social mind. Evolutionary psychology and child development. New York: The Guilford Press.
123
350
G. Grossi
Farah, M.J. 1994. Neuropsychological inference with an interactive brain: A critique of the locality assumption. Behavioral and Brain Sciences 17: 43–104. Fodor, J.A. 1983. The modularity of mind. Cambridge, MA: MIT Press. Fontanini, A., and D.B. Katz. 2008. Behavioral states, network states, and sensory response variability. Journal of Neurophysiology 100: 1160–1168. Fox, E., L. Griggs, and E. Mouchlianitis. 2007. The detection of fear-relevant stimuli: Are guns noticed as quickly as snakes? Emotion 7(4): 691–696. Friston, K.J. 1997. Imaging cognitive anatomy. Trends in Cognitive Sciences 1(1): 21–27. Friston, K.J., and C.J. Price. 2001. Dynamic representations and generative models of brain function. Brain Research Bulletin 54(3): 275–285. Gantt, E.E., B.S. Melling, and J.S. Reber. 2012. Mechanisms or metaphors? The emptiness of evolutionary psychology explanations. Theory & Psychology 22(6): 823–841. Gauthier, I., T. Curran, K.M. Curby, and D. Collins. 2003. Perceptual interference supports a nonmodular account of face processing. Nature Neuroscience 6(4): 428–432. Gauthier, I., P. Skudlarski, J.C. Gore, and A.W. Anderson. 2000. Expertise for cars and birds recruits brain areas involved in face recognition. Nature Neuroscience 3(2): 191–197. Gauthier, I., and M.J. Tarr. 2002. Unraveling mechanisms for expert object recognition: Bridging brain activity and behavior. Journal of Experimental Psychology: Human Perception and Performance 28(2): 431–446. Gauthier, I., M.J. Tarr, A.W. Anderson, P. Skudlarski, and J.C. Gore. 1999. Activation of the middle fusiform ‘face area’ increases with expertise in recognizing novel objects. Nature Neuroscience 2(6): 568–573. Ghanzanfar, A.A., and C.E. Schroeder. 2006. Is neocortex essentially multisensory? Trends in Cognitive Science 10(6): 278–285. Gick, B., and D. Derrick. 2009. Aero-tactile integration in speech perception. Nature 462(7272): 502–504. Gottlieb, G. 1998. Normally occurring environmental and behavioral influences on gene activity: From central dogma to probabilistic epigenesis. Psychological Review 105(4): 792–802. Hugdahl, K., and B.H. Johnsen. 1989. Preparedness and electrodermal fear conditioning: Ontogenetic vs phylogenetic explanations. Behavioral Research and Therapy 27: 269–278. Kanwisher, N. 2000. Domain specificity in face perception. Nature Neuroscience 3: 759–763 Kanwisher, N., and G. Yovel. 2006. The fusiform face area: A cortical region specialized for the perception of faces. Philosophical Transactions of the Royal Society B 361: 2109–2128. Karmiloff-Smith, A. 2006. Modules, genes and evolution: What have we learned from atypical development? In Attention and Performance XXI: Processes of Change in Brain and Cognitive Development, ed. Y. Munakata, and M.H. Johnson, 563–583. Oxford: Oxford University Press. Karmiloff-Smith, A. 2013. Challenging the use of adult neuropsychological models for explaining neurodevelopmental disorders: Developed versus developing brains. The Quarterly Journal of Experimental Psychology 66(1): 1–14. Jung, K., E. Ruthruff, J. Tybur, N. Gaspelin, and G. Miller. 2012. Perception of facial attractiveness requires some attentional capacity: Implications for the ‘‘automaticity’’ of psychological adaptations. Evolution and Human Behavior 33: 241–250. Laland, K.N., N. Sterelny, J. Odling-Smee, W. Hoppitt, and T. Uller. 2011. Cause and effect in biology revisited: Is Mayr’s proximate–ultimate dichotomy still useful? Science 334: 1512–1516. Lickliter, R., and H. Honeycutt 2003. Developmental dynamics and contemporary evolutionary psychology: Status quo or irreconcilable views? Reply to Bjorklund (2003), Krebs (2003), Buss and Reeve (2003), Crawford (2003), and Tooby et al. (2003). Psychological Bulletin 129(6):866–872. Lloyd, E.A. 1999. Evolutionary psychology: The burdens of proof. Biology and Philosophy 14: 211–233. Marr, D. 1982/2010. Vision: A computational investigation into the human representation and processing of visual information. Cambridge (MA): MIT Press. Masataka, N. 1993. Effects of experience with live insects on the development of fear of snakes in squirrel monkeys, Saimiri sciureus. Animal Behavior 46: 741–746. McIntosh, A.R. 2000. Towards a network theory of cognition. Neural Networks 13: 861–870. McGurk, H., and J. MacDonald. 1976. Hearing lips and seeing voices. Nature 264: 746–748. McKeeff, T.J., R.W. McGugin, F. Tong, and I. Gauthier. 2010. Expertise increases the functional overlap between face and object perception. Cognition 117(3): 355–360. Mineka, S., M. Davidson, M. Cook, and R. Keir. 1984. Observational conditioning of snake fear in rhesus monkey. Journal of Abnormal Psychology 93(4): 355–372.
123
A module is a module
351
¨ hman. 2002. Phobias and preparedness: The selective, automatic, and encapsulated Mineka, S., and A. O nature of fear. Biological Psychiatry 52: 927–937. Neville, H.J., and D.L. Mills. 1997. Epigenesis of language. Mental Retardation and Developmental Disabilities Research Reviews 3: 282–292. New, J., L. Cosmides, and J. Tooby. 2007. Category-specific attention for animals reflects ancestral priorities, not expertise. Proceedings of the National Academy of Sciences 104(42): 16593–16603. Oyama, S. 2000. The ontogeny of information: Developmental systems and evolution, 2nd ed. Durham, N.C.: Duke University Press. Picton, T.W., S. Bentin, P. Berg, E. Donchin, S.A. Hillyard, S.A. Johnson, and M.J. Taylor. 2000. Guidelines for using human event-related potentials to study cognition: Recording standards and publication criteria. Psychophysiology 37: 127–152. Pinker, S. 1997. How the mind works. New York: W. W. Norton & Company. Price, C.J., and K.J. Friston. 2005. Functional ontologies for cognition: The systematic definition of structure and function. Cognitive Neuropsychology 22(3/4): 262–275. Price, C., G. Thierry, and T. Griffiths. 2005. Speech-specific auditory processing specific: Where is it? Trends in Cognitive Science 9(6): 271–276. Richardson, R.C. 2007. Evolutionary psychology as maladapted psychology. Cambridge: MIT Press. Sperber, D. 1994. The modularity of thought and the epidemiology of representations. In Mapping the mind: Domain specificity in cognition and culture, ed. L.A. Hirschfeld and S.A. Gelman, 39–67. New York: Cambridge University Press. Shapiro, L., and W. Epstein. 1998. Evolutionary theory meets cognitive psychology: A more selective perspective. Mind and Language 13(2): 171–194. Stone, V.E., L. Cosmides, J. Tooby, N. Kroll, and R.T. Knight. 2002. Selective impairment of reasoning about social exchange in a patient with bilateral limbic system damage. Proceedings of the National Academy of Science 99(17): 11531–11536. Takahashi, H., M. Matsuura, N. Yahata, M. Koeda, T. Suhara, and Y. Okubo. 2006. Men and women show distinct brain activations during imagery of sexual and emotional infidelity. Neuroimage 32(3): 1299–1307. Tipples, J., A.W. Young, P. Quinlan, P. Broks, and A.W. Ellis. 2002. Searching for threat. The Quarterly Journal of Experimental Psychology 55A(3): 1007–1026. Tooby, J., and L. Cosmides. 1992. The psychological foundations of culture. In The adapted mind: Evolutionary psychology and the generation of culture, ed. J. Barkow, L. Cosmides, and J. Tooby, 19–136. New York: Oxford University Press. Tooby, J., and L. Cosmides. 2005. Conceptual foundations of evolutionary psychology. In The handbook of evolutionary psychology, ed. D.M. Buss, 5–67. Hoboken (NJ): Wiley. Vargha-Khadem, F., K. Watkins, K. Alcock, P. Fletcher, and R. Passingham. 1995. Praxic and nonverbal cognitive deficits in a large family with a genetically transmitted speech and language disorder. Proceedings of the National Academy of Sciences of the United States of America 92(3): 930–933. Westermann, G., D. Mareschal, M.H. Johnson, S. Sirois, M.W. Spratling, and M.S.C. Thomas. 2007. Neuroconstructivism. Developmental Science 10(1): 75–83. White, A.E., D.T. Kenrick, and S.L. Neuberg. 2013. Beauty at the ballot box: Disease threats predict preferences for physically attractive leaders. Psychological Science 24(12): 2429–2436. Wojciulik, E., N. Kanwisher, and J. Driver. 1998. Covert visual attention modulates face-specific activity in the human fusiform gyrus: fMRI study. Journal of Neurophysiology 79: 1574–1578.
123