Pere Alberch’s Developmental Morphospaces
and the Evolution of Cognition
Departament de Filologia Catalana and Centre
orica, Facultat de Lletres
onoma de Barcelona
Bellaterra (Barcelona), Spain
Departamento de Filolog´
nola, Facultad de Filolog´
Campus El Mil´
an, Universidad de Oviedo, Oviedo, Spain
In this article we argue for an extension of Pere Alberch’s
notion of developmental morphospace into the realm of cog-
nition and introduce the notion of cognitive phenotype as a new
tool for the evolutionary and developmental study of cognitive
computational complexity, evolutionary developmental
biology (EvoDevo), evolution of cognition, language,
March 13, 2009; revised and accepted June 10, 2009
Biological Theory 3(4) 2008, 297–304. c
2009 Konrad Lorenz Institute for Evolution and Cognition Research 297
Pere Alberch’s Developmental Morphospaces and the Evolution of Cognition
During the 1980s and until the time of his death in the late
1990s, the Catalan biologist Pere Alberch developed a theo-
retical model of morphological evolution. Alberch’s model can
in many ways be considered as a direct precursor of contempo-
rary evolutionary developmental biology (EvoDevo) thinking
(for details and an historical overview see Garc´
2005; Etxeberria and Nu˜
no de la Rosa in press). As we ex-
pound in detail in the next section, one of the characteristic
features of Alberch’s model is the idea that “morphological
variation at the macroscopic level is not continuously dis-
tributed. Rather it is distributed among a ﬁnite set of discrete
states” (Alberch 1980: 654); moreover, it isalso nonchaotic, in
the sense that “morphologies are not generated in a continuous
and random fashion” (p. 654). This idea goes back at least to
the work of 19th-century transcendental morphologists like ´
Geoffroy Saint-Hilaire (1818–1822), I. Geoffroy Saint-Hilaire
(1832–1837), and Owen (1848) as well as to that of some post-
Darwinian naturalists like Bateson (1894); and it was directly
implemented by Alberch in his notion of parametric space.
For the purposes of these introductory remarks, sufﬁce it to
say that a parametric space is a theoretical representation of
the ﬁnite collection of possible phenotypes as a function of a
set of developmental constraints or morphogenetic parameters
that is also ﬁnite. In Alberch’s view, such spaces are to be seen
not as static but as dynamic entities, where possible transitions
from one phenotype to another may be deﬁned as perturba-
tions over one or more of these morphogenetic parameters,
thus making some transitions more probable than others and,
concomitantly, some phenotypes more probable than others.
Although Alberch’s own experimental work concentrated
on the study of the development of some very speciﬁc morpho-
logical patterns (like the number of digits per limb in different
organisms), his model was intended to explain any case of
morphological variation, including that of the nervous sys-
tem. Indeed, Alberch’s theories are particularly well suited to
any attempt at explaining brain development and evolution,
since, as most contemporary work in this ﬁeld suggests, these
processes are clearly nonrandom (Striedter 2005, 2006; see
also Ebbesson 1980, 1984 and Edelman 1987 for some earlier
proposals) and constrained to a ﬁnite number of possibilities
(Hofman 2001). Now, inasmuch as the “mind” is just what “the
brain does” (Searle 1985), this line of research should prove
also very useful in the enterprise of broadening some EvoDevo
concepts to aid in explaining the evolution of cognition, follow-
ing an internalist trend that is gaining visibility (e.g., Grifﬁths
and Stotz 2000; Amundson 2006; Grifﬁths 2007) as an alterna-
tive to the externalist positioning of evolutionary psychology
(Barkow et al. 1992; Pinker 1997; Plotkin 1997; Buss 2005,
A sizeable literature is already devoted to explaining the
evolution of patterns of neuroanatomical organization based
on the idea that readjustments in the development of ner-
vous systems are the main evolutionary source of diversity
in this domain (see Parker et al. 2000, Falk and Gibson 2001,
and Minugh-Purvis and McNamara 2001 on primate brain
evolution). While this is an important factor linking standard
EvoDevo applications to the study of cognition, we argue that
it cannot be seen as the ultimate move in explaining the origins
of complex forms of cognition. This is so because perturba-
tions leading to minor quantitatively measurable changes in
brain organization can nevertheless correlate with true quali-
tative gaps in cognitive skills. For this reason we defend the
position that the adoption of an abstract and function-oriented
viewpoint is an important complementary strategy to that of
a purely anatomical perspective when dealing with the evo-
lutionary study of cognition (on the issue of the levels of
brain description, see Chomsky 1980; Marr 1982), a view-
point that can nicely be accommodated within EvoDevo en-
deavors (see Gibson 1990, 2004, Parker and McKinney 1999,
and Langer 2000 for some precedents of the idea in various
cognitive areas). Actually, Alberch himself explored these av-
enues in his notebooks, where we ﬁnd some reﬂections on
the pattern-generating capacities of the human brain, invok-
ing the computer metaphor (Laura Nu˜
no de la Rosa, personal
communication, 11 March 2008).
This article presents an extension of Alberch’s notion of
parametric space into the realm of cognition, suggesting that
a parametric space of cognitive or computational phenotypes
exists, paralleling the space of possible brain morphologies,
such that each computational phenotype correlates with one
or more brain phenotypes of which it is a computational char-
acterization. The space of possible computational phenotypes
is deﬁned on the basis of the Chomsky hierarchy of computa-
tional levels of complexity (Chomsky 1956, 1959), for which
a precise mathematical characterization is available in terms
of computational regimes and resources. Thus, we have at our
disposal a theoretical tool to provide a functional description
of the necessary resources to perform some cognitive task, ca-
pable of complementing and possibly directing research at the
level of brain morphology. At the end of this article we offer a
concrete example of this application to the case of the human
faculty of language.
Morphological Evolution and Related Concepts
Alberch (in particular 1980, 1989, and 1991) originated the
concept of morphological evolution aimed at explaining the
phylogeny of organic designs on the very same basis as modern
EvoDevo thinking: changes in the parameters underlying the
development of organisms are the main source of the evolution-
ary processes capable of introducing novelties into nature. He
also contended that developmental systems possess the prop-
erties of complex dynamic systems (Thelen and Smith 1994;
Kelso 1995), in which intricate interactions between genetic
298 Biological Theory 3(4) 2008
Sergio Balari and Guillermo Lorenzo
and nongenetic factors relate nonlinearly with morphological
outputs. Such nonlinearity basically means that once certain
critical thresholds are reached, small perturbations of any of
the morphogenetic parameters of the system are capable of
bringing about wide-range consequences in development and,
ultimately, in the evolution of entire lineages of organisms. In
this connection, De Renzi et al. (1999: 625–626) have pointed
out that Alberch anticipated the application of contemporary
complex systems theory to organic development through his
conception of developmental systems as dynamic nonlinear
systems (see, in particular, Oster and Alberch 1982; see also
Kauffman 1993, which includes a note of personal acknowl-
edgment and several references to Alberch’s work). As a mat-
ter of fact, at the time of his death Alberch was working on
the project of formalizing his ideas in terms of chaos and
complexity theory in a book titled An Introduction to Chaos
Theory and Complexity With Special Emphasis on Biological
Sciences, which he left unﬁnished (L. Nu˜
no de la Rosa, per-
sonal communication, 11 March 2008). Alberch also shared
with most EvoDevo theoreticians the conviction that the most
common effects of those perturbations have to do with the tim-
ing and/or rates of growth of the developing structures, which
as a result can diverge radically from closely related ones
(Gould 1977; Smith 2001, 2002). Heterochrony was thus the
main mechanism referred to by Alberch to explain the origins
of most evolutionary novelties (Alberch et al. 1979; Alberch
and Alberch 1981; see also Alberch 1985 and Alberch and
Blanco 1996 for some critical comments on standard uses of
A more distinctive aspect of Alberch’s model is the idea
that systems of interactions underlying developmental se-
quences are rather stable and that changes in these sequences
are mostly due to modiﬁcations in the values of one or another
of the morphogenetic parameters of the system (Alberch 1989,
1991). In the study of complex dynamic systems, the concept
of control parameter refers to the systemic component whose
perturbations correlate with the emergence of new morpholo-
gies (a new pattern in the surface of a chemical solution, a new
embryological state, a new form of behavior, and so on; Thelen
and Smith 1994: 63–64). A control parameter is not, however,
a central agent in the causation of phenotypic variation in that
the effects of its perturbations do not immediately reﬂect on
morphological outputs but on the other morphogenetic param-
eters (Thelen and Smith 1994: 112; Kelso 1995: 7). Within this
model, development (even the development of novel forms) is
always a function of the system as a whole. The idea of control
parameter basically introduces the possibility of pinpointing
a single parameter of the system as the starting point of the
chain reaction leading to new morphologies.
For our own purposes, however, the most relevant aspect
of Alberch’s proposals is the contention that developmental
systems foreshadow the scope of their attainable phenotypes,
An idealized parametric space deﬁned by two abstract parameters xand y.
as well as the trajectories leading from one phenotypic state
to another. It is the concept of parametric space that in Al-
berch’s framework is in charge of theoretically representing
the ﬁnite and discrete set of the possible outcomes of any
developmental system (Alberch 1989, 1991). The main prop-
erties of parametric spaces are summed up in the following
paragraphs (see Figure 1 as a point of reference).
A parametric space is a ﬁnite set of discrete phenotypes.
The discontinuous character of phenotypic variation is cap-
tured in Figure 1 by the spaces labeled with capital letters.
Each phenotype has a characteristic probability of coming
into being, represented in Figure 1 by the extension that it
occupies (D is thus the most probable phenotype, while B is
the most improbable one). Moreover, each phenotype is also
characterized by the relative probability of transforming itself
into one or another of the neighboring phenotypes. In Figure 1
this aspect is represented by the extension of the line separat-
ing different phenotypes (A has a strong probability of turning
into D, a low probability of turning into B, and no probability
at all of turning into any of the remaining phenotypes).
Species are represented in Figure 1 by means of the oval
items (s1and s2). From a population point of view, the model
incorporates the following contentions: Firstly, every species
falls upon one or another phenotype (s1belongs to phenotype
D, whereas s2ﬁts in phenotype A). Secondly, the morphologi-
cal stability of a species is a function of both (1) the probability
of its phenotype (s1is, in principle, a more stable population
than s2) and (2) its proximity to a point of bifurcation to other
phenotypes (s1is thus a rather unstable population within its
phenotype, given its vicinity to the bifurcation leading to E and
F). Finally, the proximity to a point of bifurcation as well as
the relative propensity of its own phenotype to transform into
one or another phenotype puts a certain population at the edge
of undergoing a radical morphological reorganization (s1,for
instance, has a high propensity of acquiring the properties of
Biological Theory 3(4) 2008 299
Pere Alberch’s Developmental Morphospaces and the Evolution of Cognition
phenotype E). Bifurcation, a point within a parametric space
in which a minimal perturbation is capable of bringing about
qualitatively new morphologies, is thus another key concept
of Alberch’s framework (Oster and Alberch 1982).
From an evolutionary perspective, the idea that the geom-
etry of parametric spaces works as a very strong constraining
force, capable of counteracting that of natural selection, is an
important corollary of this model. As explicitly stated by Al-
berch, this does not discredit the role of natural selection in
the evolution of organic designs; instead, it redeﬁnes it as a
ﬁltering mechanism rather than as a creative force (Alberch
1980: 664, 1989: 46–48, 1991: 16; Oster and Alberch 1982:
455; see also Goodwin 1994: 143; Wagensberg 2004: 125).
As we argue in the next section, Alberch’s model of mor-
phological evolution can be easily extended to explain the
evolution of complex cognitive functions.
As we noted in the introductory section, a detailed character-
ization of all the deﬁning properties of a speciﬁc brain phe-
notype is not always possible, but we have at our disposal
a powerful theoretical tool that makes it possible to identify
its main features at the computational level and to elaborate
concrete proposals about what structures and neural organiza-
tions could be associated with these properties. Our concrete
proposal is to assume the existence of a parametric space of
cognitive phenotypes containing a minimum of four pheno-
types, in direct correspondence with the four levels of compu-
tational complexity in the Chomsky hierarchy, namely, type 3
(regular), type 2 (context-free), type 1 (context-sensitive), and
type 0 (unrestricted). For ease of exposition, we limit our space
to four phenotypes, but it should be taken into account that
after Chomsky’s original formulation, the complexity space
has been enriched to accommodate further levels (Chomsky
1963; Aho 1968; Hopcroft and Ullman 1979: ch. 14; Joshi
1985; Joshi et al. 1991; Vijay-Shanker and Weir 1994). Since
nothing in this exposition hinges on the exact number of phe-
notypes, we stick to a space of four, which, following the
numbering in the hierarchy, we will label CP3,CP
CP0, respectively. Just as in Figure 1 above, these would be
arranged in a parametric space, representing the probability of
each phenotype and the probability of jumping from one to
the others. The only exception here will be CP0because we
assume that the probability of transition from any other phe-
notype to CP0is actually 0. In this sense, our parametric space
is a theoretical morphospace in the sense of Rasskin-Gutman
(2005: 214–215), including possible (both actual and poten-
tial) and impossible phenotypes, as opposed to an empirical
morphospace, excluding the latter. Its exclusion is justiﬁed
by natural limitations on the parametric factors involved, as it
seems reasonable to do with the case of CP0, given the proper-
ties associated with unrestricted systems (see Chomsky 1959
for discussion of this point).
What is important here is that whatever the speciﬁc mor-
phological properties of some brain phenotype, but provided
that we are able to associate it with one computational pheno-
type, we have a precise computational characterization for it.
In fact, we know that CP3has a computational power equiv-
alent to that of a ﬁnite-state automaton, CP2is equivalent to
a pushdown automaton, CP1to a linear-bounded automaton,
and, ﬁnally, CP0is equivalent to a Turing machine.
Space limitations prevent us from giving a precise for-
mal description of each of these systems, but they can be
safely understood as kinds of abstract computers working with
speciﬁc processing regimes and with different computational
resources. This sketchy characterization is nevertheless sufﬁ-
cient to derive an interesting conclusion with potentially rel-
evant repercussions in the study of the evolution of cognitive
abilities. In fact, as a number of recent mathematical results
demonstrate (Weir 1992, 1994; Joshi and Schabes 1997), the
only difference between the different levels of complexity lies
in a single computational resource, viz., memory, meaning
that the progression up the scale of complexity is simply a
function of the changes introduced in the memory system,
with no other modiﬁcation of any fundamental property of the
computational system being necessary. We are, therefore, in
a position not just of being able to characterize precisely our
computational phenotypes but also of being able to determine
the kinds of alterations of the developmental system that are
necessary to facilitate the transition from one phenotype to
the other and, concomitantly, eventually to characterize the
alterations at the morphological level necessary to implement
one or another computational regime. The obvious corollary
of this is that the evolution of higher cognitive skills may be
understood as a historical process in which newer and more
sophisticated cognitive skills have emerged not as the result of
radical rearrangements in the computational regimes of ner-
vous systems but just as the result of brains having at their
disposal larger and more sophisticated systems of memory.
To conclude this section, note, however, that the paramet-
ric space of cognitive computational phenotypes possesses the
very same properties as the morphospaces that Alberch pro-
posed. That is, it shows a noncontinuous distribution of phe-
notypic variation, with discrete and easily identiﬁable states.
Thus, the transition from one state to another is, in fact, a
“jump” that is only made possible once a speciﬁc critical point
is attained (possibly as a result of the accumulation of small
gradual changes). In this connection, it is nevertheless impor-
tant to take into account that the fact that there is discontinuity
in variation does not contradict the idea of gradual change at
the level of processes (as already pointed out by Bateson 1894:
13–17). This is, in fact, a fundamental feature of the concept
of “critical-point emergence” to which we shall appeal here
300 Biological Theory 3(4) 2008
Sergio Balari and Guillermo Lorenzo
(for details, see Reid 2007: ch. 8). According to our proposal,
then, the evolution of what we call “the computational mind”
would have consisted in a historical process in which complex
interactions among genetic and epigenetic factors during the
individual developmental process of the nervous system would
have given rise to qualitatively differentiated phenotypes via
a system of “critical-point” emergent processes. Such pheno-
types would have been able to act as the material support for
richer and varied computational regimes, as we will see in the
Some Examples: Language, Birdsong, and Knots
Language is often seen “as one of the chief distinctions be-
tween man and the lower animals,” as Darwin (1871) put it.
Language is complex—but not too complex. In fact, as several
years of investigations of the structural complexities of human
language have demonstrated, the computational resources nec-
essary to process the most structurally complex human utter-
ance possible appear to be only slightly beyond the capacity
of a context-free system (our CP2) and within the complex-
ity space of what Joshi (1985) termed mild context-sensitivity
(i.e., within the lower area of our CP1) but in any case, never
beyond what are known as indexed systems (see Aho 1968,
Pullum 1986, and Radzinski 1991 for critical discussion).
That said, and given our previous considerations on the
role of memory in connection to greater computational com-
plexity, we can identify the major evolutionary event that gave
rise to modern human linguistic abilities—or, to be precise, the
computational system underlying the faculty of language in the
narrow sense of Hauser et al. (2002)—as one that enhanced
the working memory of a preexisting computational system
in the direction of context-sensitivity (CP1). In fact, follow-
ing Lieberman (2006), there is enough evidence for such an
anatomical precursor in the form of the cortico-subcortico-
cortical circuits associated with the regulation of different as-
pects of mobility, cognition, and emotion whose main sub-
cortical component is located in the basal ganglia (see also
Cummings 1993). In this model the role of the cortical areas is
that of working memory space, with the basal ganglia acting
as a cognitive pattern generator to make up what Lieberman
characterizes as an iterative sequencing machine. From an
evolutionary point of view the basal ganglia, as opposed to the
cortex, have been described as highly conservative structures
among amniotes (Reiner et al. 1984), such that no greater inter-
speciﬁc differences are expected at the level of the subcortical
structure acting as the material support of the system’s pattern-
generation procedure, leaving all room for innovation to those
cortical areas that, as pointed out above, provide working space
memory to the system.
This is all very sketchy (but see Balari and Lorenzo 2009
for a concrete and detailed proposal of the emergence of lan-
guage along these lines); however, we believe it is illustrative
enough of the applicability of our notion of computational
phenotype in the construction of scientiﬁc hypotheses about
the evolution and development of higher cognitive functions
in humans and other species. Indeed, we believe that one of
the virtues of our proposal is that it can easily be applied to
the cognitive abilities of other species.
An important claim of our proposal is that the compu-
tational core engine of the vertebrate brain is located in the
striatum, with the basal ganglia functioning as an iterative se-
quencing machine. The participation of the basal ganglia in
the performance of linguistic tasks appears to be unquestion-
able (Tettamanti et al. 2005), and more interestingly, they also
appear to play a crucial role in singing behavior in birds, with
striking neurochemical similarities (Ding and Perkel 2002;
Gale and Perkel 2005; Sasaki et al. 2006; Cornil et al. 2008;
Huang and Hessler 2008). In addition, at the genetic level,
the human variety of FOXP2 is expressed, among many other
loci, in the basal ganglia (see Ben´
ıtez Burraco 2008, 2009
for a detailed overview), as it is the case of its avian ho-
molog (Rochefort et al. 2007). All this evidence suggests that
birdsong and language share a common neural substrate and,
consequently, that the observed differences between the two
cognitive abilities must be due to other properties of the com-
putational systems associated with them. One of these differ-
ences, according to the proposal developed here, must lie at
the level of the resources available to perform these tasks. In
fact, we already noted that, computationally, language would
correspond to our CP1, whereas birdsong typically shows the
properties of CP3(Todt and Hultsch 1998; Okanoya 2002),
and in any case, it never appears to go beyond those of CP2
(Gentner et al. 2006). Clearly, however, this cannot be the
only difference because language is also a symbolic system
that precisely because it has speciﬁc computational properties,
also possesses the property of discrete inﬁnity. This is impor-
tant because, as pointed out by Lorenzo (2006), it is perfectly
plausible to assume that in the animal kingdom certain be-
haviors are observed that suggest the presence of a symbolic
system (or the rudiments thereof) or the use of very complex
recursive patterns (which does not imply the simultaneous
presence of complex recursion and symbolism). To extend and
perfect working memory space is, then, a necessary condition
for the emergence of language but not a sufﬁcient one. This
dissolves a potential objection to our notion of computational
phenotype, viz., that it may not necessarily be the case that
this phenotype (and its corresponding morphological pheno-
type) correlates with the presence of language. In fact, this
observation is correct, and our CP1might well correspond
to nonlinguistic “minds” that would nevertheless be capable
of producing complex recursive patterns within other areas
of cognition such as motor sequences or melodic sequences.
Remember that, according to Lieberman’s (2006) model, the
Biological Theory 3(4) 2008 301
Pere Alberch’s Developmental Morphospaces and the Evolution of Cognition
basal ganglia comprise a sequence of cognitive patterns, but
cognitive patterns may be of many different sorts, and the
sequencer, just because of connectivity and working space
limitations, might have access to only a single type of pattern
or to a reduced collection of them (e.g., motor patterns and
melodic patterns) but not others (symbolic patterns, assuming
these are even available). In this connection, it is particularly
interesting to pay attention to, for instance, the construction
behavior of weaver birds, which are capable of making some
types of knots (Hansell 2000, 2005). Making a knot requires
the application of an operation over a part of the constructed
ﬁgure and keeping it in active memory until the moment at
which the operation completing the ﬁgure is executed, which
requires complex computations perhaps within the power of
CP1(Camps and Uriagereka 2006).
We need to emphasize that the computational system to
which we refer must not be identiﬁed with the faculty of lan-
guage in the broadest sense of the term (Hauser et al. 2002).
Moreover, since our model does not presuppose that this sys-
tem is language-speciﬁc (or human-speciﬁc, for that matter), it
may perfectly be understood as one of the “infrastructural” el-
ements (Oller 2005) rather than as a “component” of language
per se. In this sense, this system would correspond to a level of
cognitive analysis comparable to other “facilitating” elements
of the linguistic capacity, such as shared attention (Carpenter
et al. 1998; Tomasello and Farrar 1986) or vocal dexterity
(Ploog 2002; Oller 2005). We wish to point out in this connec-
tion that our model offers the possibility of explaining these
infrastructural elements of the language faculty, invoking the
very same factors from which we derive the complexity of its
computational regime. On the one hand, it is known that con-
trolled attention, of which shared attention can be assumed to
be just a particular kind, may be enhanced as a function of the
working-memory resources associated with it (see Desimone
1996; Awh and Jonides 2001; Curtis and D’Esposito 2003; see
also the overview in Klingberg 2009). On the other hand, vocal
dexterity is explained by Striedter (2005: 324–326) as an effect
of the “connectional invasion” of the medulla and the spinal
cord by cortical neurons as a by-product of cortical growth.
We believe that these considerations have important
methodological implications, especially with respect to the ap-
plication of the comparative method when seeking precursors
of language or of some of its constituent properties like com-
plex recursion. Indeed, one of the most direct consequences
is that formal grammar and automata theory may prove to be
an extremely useful tool at the time of assessing the abilities
and capabilities shown by other animal species, a point also
argued for by many other researchers (see, e.g., Hauser et al.
2002; Fitch and Hauser 2004; O’Donnell et al. 2005; Camps
and Uriagereka 2006; Gentner et al. 2006; Uriagereka 2008).
However, as we have just seen, there is no guarantee that animal
species different from ours show the very same abilities within
the very same areas of cognition, which is perfectly compati-
ble with our suggestion that a computational phenotype may
correlate with many different morphological phenotypes cor-
responding to different cognitive abilities. Thus, our proposal
is less prone to be the target of the kind of criticisms (Perruchet
and Rey 2005; Pullum and Rogers 2006; Rogers and Pullum
in press) that were targeted at work along similar lines. Fi-
nally, just as our proposal is compatible with the possibility of
nonlinguistic expert minds associated with our CP1—which
would, for example, explain the absence of any clear evi-
dence of symbolic behavior in the Neandertal archaeological
record (Balari et al. 2008; Ben´
ıtez Burraco et al. 2008)—it is
also compatible with the idea that the computational regime
subserving the human language faculty also subserves other,
nonlinguistic aspects of human mentality, and, consequently
that the narrow faculty of language of Hauser et al. (2002) is,
in fact, not speciﬁc to language.
Our work may be inscribed within a line of research charac-
terized by the appeal to both internalist principles and compu-
tational complexity. We believe that, pace Lewontin’s (1998)
pessimistic stance with respect to the evolutionary study of
cognition, attempts to construct evolutionary explanations of
complex cognitive abilities will eventually open new pathways
to ﬁnd empirical support for Darwin’s (1871) contention that
“nevertheless the difference in mind between man and the
higher animals, great as it is, certainly is one of degree and not
We wish to thank the editor-in-chief of Biological Theory and two anonymous
reviewers for their useful suggestions, which helped improve the contents of
this article. All remaining errors are our own. We are indebted to Laura Nu˜
de la Rosa for putting Pere Alberch’s notebooks at our disposal. This work
has been carried out through the project “Bioling¨
ıstica: fundamento gen´
desarrollo y evoluci´
on del lenguaje” (HUM2007-60427/FILO) of the Spanish
Ministerio de Educaci´
on y Ciencia, which was partially cofunded by FEDER
funds (Balari and Lorenzo) and received support from the Generalitat de
Catalunya through grant 2005SGR 00753 Ling¨
orica to the Centre
orica of the Universitat Aut`
onoma de Barcelona (Balari).
Aho AV (1968) Indexed grammars: An extension of context-free grammars.
Journal of the ACM 15: 647–671.
Alberch P (1980) Ontogenesis and morphological diversiﬁcation. American
Zoologist 20: 653–667.
Alberch P (1985) Problems with the interpretation of developmental se-
quences. Systematic Zoology 34: 46–58.
Alberch P (1989) The logic of monsters: Evidence for internal constraint in
development and evolution. Geobios 12 (m´
Alberch P (1991) Del gen al fenotipo: sistemas din´
amicos y evoluci´
ogica. Revista Espa˜
nola de Paleontolog´
umero extraordinario “El
302 Biological Theory 3(4) 2008
Sergio Balari and Guillermo Lorenzo
estudio de la forma org´
anica y sus consecuencias en Paleontolog´
ıa y Paleontolog´
ıa Evolutiva”): 13–19.
Alberch P, Alberch J (1981) Heterochronic mechanisms of morphological di-
versiﬁcation and evolutionary change in the neotropical salamander, Boli-
toglossa occidentalis (Amphibia: Plethodontidae). Journal of Morphology
Alberch P, Blanco MJ (1996) Evolutionary patterns in ontogenetic transfor-
mation. International Journal of Developmental Biology 40: 845–858.
Alberch P, Gould SJ, Oster GF, Wake DB (1979) Size and shape in ontogeny
and phylogeny. Paleobiology 5: 296–317.
Amundson RA (2006) EvoDevo as cognitive psychology. Biological Theory
Awh E, Jonides J (2001) Overlapping mechanisms of attention and spatial
working memory. Trends in Cognitive Science 5: 119–126.
Balari S, Ben´
ıtez Burraco A, Camps M, Longa VM, Lorenzo G, Uriagereka
J (2008) ¿Homo loquens neanderthalensis? En torno a las capaci-
olicas y ling¨
ısticas del Neandertal. Munibe Antropologia—
Arkeologia 59: 3–24.
Balari S, Lorenzo G (2009) Computational phenotypes: Where the theory of
computation meets Evo-Devo. Biolinguistics 3: 2–60.
Barkow JH, Cosmides L, Tooby J, eds (1992) The Adapted Mind: Evolutionary
Psychology and the Generation of Culture. Oxford: Oxford University
Bateson W (1894) Materials for the Study of Variation Treated with Special
Regard to Discontinuity in the Origin of Species. London: Macmillan.
ıtez Burraco A (2008) FOXP2 y la biolog´
ıa molecular del lenguaje: Nuevas
evidencias. II. Aspectos moleculares e implicaciones para la ontogenia y
la ﬁlogenia del lenguaje. Revista de Neurolog´
ıa 46: 351–359.
ıtez Burraco A (2009) Genes y Lenguaje: Aspectos ontogen´
eticos y cognitivos. Barcelona: Revert´
ıtez Burraco A, Longa VM, Lorenzo G, Uriagereka J (2008) Also sprach
Neanderthalis. . . Or did she? Biolinguistics 2: 225–232.
Buss DM, ed (2005) The Handbook of Evolutionary Psychology. Hoboken,
Buss DM (2007) Evolutionary Biology: The New Science of the Mind. Boston:
Allyn and Bacon.
Camps M, Uriagereka J (2006) The Gordian knot of linguistic fossils. In: The
Biolinguistic Turn: Issues on Language and Biology (Rossell´
J, eds), 34–65. Barcelona: PPU.
Carpenter M, Nagell K, Tomasello M, Butterworth G, Moore C (1998) Social
cognition, joint attention, and communicative competence from 9 to 15
months of age. Monographs of the Society for Research in Child Devel-
opment 63: 1–174.
Chomsky N (1956) Three models for the description of language. IRE Trans-
actions on Information Theory 2: 113–124.
Chomsky N (1959) On certain formal properties of grammars. Information
and Control 2: 137–167.
Chomsky N (1963) Formal properties of grammars. In: Handbook of Mathe-
matical Psychology, Vol. 2 (Luce RD, Bush RR, Galanter E, eds),323–418.
New York: Wiley.
Chomsky N (1980) Rules and Representations. New York: Columbia Univer-
Cornil CA, Castelino CB, Ball GF (2008) Dopamine binds to α2-adrenergic re-
ceptors in the song control of zebra ﬁnches (Taeniopygia guttata). Journal
of Chemical Neuroanatomy 35: 202–215.
Cummings JL (1993) Frontal–subcortical circuits and human behavior.
Archives of Neurology 50: 873–880.
Curtis CE, D’Esposito M (2003) Persistent activity in the prefrontal cortex
during working memory. Trends in Cognitive Science 7: 415–423.
Darwin C (1871) The Descent of Man, and Selection in Relation to Sex.
London: John Murray.
o J (1999) Obituary. Evolution, development and
complexity in Pere Alberch (1954–1998). Journal of Evolutionary Biology
Desimone R (1996) Neural mechanisms for visual memory and their role
in attention. Proceedings of the National Academy of Science USA 93:
Ding L, Perkel DJ (2002) Dopamine modulates excitability of spiny neurons
in the avial basal ganglia. Journal of Neuroscience 22: 5210–5218.
Ebbesson SOE (1980) The parcellation theory and its relation to interspeciﬁc
variability in brain organization, evolutionary and ontogenetic develop-
ment and neuronal plasticity. Cell and Tissue Research 213: 179–212.
Ebbesson SOE (1984) Evolution and ontogeny of neural circuits. Behavioral
and Brain Sciences 7: 321–366.
Edelman GM (1987) Neural Darwinism: The Theory of Neuronal Group
Selection. New York: Basic Books.
Etxeberria A, Nu˜
no de la Rosa L (in press) A world of opportunity within
constraint: Pere Alberch’s early Evo-Devo. In Pere Alberch: The Creative
Trajectory of an Evo-Devo Biologist (Rasskin-Gutman D, De Renzi M,
encia: Universitat de Val`
Falk D, Gibson KR, eds (2001) Evolutionary Anatomy of the Primate Cerebral
Cortex. Cambridge: Cambridge University Press.
Fitch WT, Hauser MD (2004) Computational constraints on syntactic pro-
cessing in a nonhuman primate. Science 303: 377–380.
Gale SD, Perkel DJ (2005) Properties of dopamine release and uptake in the
songbird basal ganglia. Journal of Neurophysiology 93: 1871–1879.
ıa-Azkonobieta T (2005) Evoluci´
on, desarrollo y (auto)organizaci´
estudio sobre los principios ﬁlos´
oﬁcos de la evo-devo. Doctoral Disserta-
tion, Universidad del Pa´
ıs Vasco, Donostia, Spain.
Gentner TQ, Fenn KM, Margoliash D, Nusbaum H (2006) Recursive syntactic
pattern learning by songbirds. Nature 440: 1204–1207.
Geoffroy Saint-Hilaire ´
E (1818–1822) Philosophie anatomique, 2 Vols. Paris:
Geoffroy Saint-Hilaire I (1832–1837) Histoire g´
erale et particuli`
anomalies de l’organisation chez l’homme et les animaux, 4 Vols. Paris:
Gibson KR (1990) New perspectives on instincts and intelligence: Brain size
and the emergence of hierarchical mental construction skills. In: “Lan-
guage” and Intelligence in Monkeys and Apes (Parker ST, Gibson KR,
eds), 97–128. New York: Cambridge University Press.
Gibson KR (2004) Human brain evolution: Developmental perspectives. In:
Biology and Knowledge Revisited: From Neurogenesis to Psychogenesis
(Parker ST, Langer J, Milbrath C, eds), 123–143. Mahwah, NJ: Erlbaum.
Goodwin B (1994) How the Leopard Changed Its Spots: The Evolution of
Complexity. London: Phoenix.
Gould SJ (1977) Ontogeny and Phylogeny. Cambridge, MA: Belknap Press.
Grifﬁths PE (2007) Evo-Devo meets the mind:Towards a developmental evo-
lutionary psychology. In: Integrating Evolution and Development (Bran-
don R, Sansom R, eds), 195–226. Cambridge, MA: MIT Press.
Grifﬁths PE, Stotz K (2000) How the mind grows: A developmental perspec-
tive on the biology of cognition. Synthese 122: 29–51.
Hansell MH (2000) Bird Nests and Construction Behaviour.Cambridge: Cam-
bridge University Press.
Hansell MH (2005) Animal Architecture. Oxford: Oxford University Press.
Hauser MD, Chomsky N, Fitch WT (2002) The faculty of language:
What is it, who has it, and how did it evolve? Science 298: 1569–
Hofman MA (2001) Brain evolution in hominids: Are we at the end of the
road? In: Evolutionary Anatomy of the Primate Cerebral Cortex (Falk D,
Gibson KR, eds), 113–127. Cambridge: Cambridge University Press.
Hopcroft JE, Ullman JD (1979) Introduction to Automata Theory, Languages,
and Computation. Reading, MA: Addison-Wesley.
Biological Theory 3(4) 2008 303
Pere Alberch’s Developmental Morphospaces and the Evolution of Cognition
Huang YC, Hessler NA (2008) Social modulation during songbird courtship
potentiates midbrain dopaminergic neurons. PloS ONE 3(10): e3281.
Joshi AK (1985) Tree adjoining grammars: How much context-sensitivity is
required to provide reasonable structural descriptions? In: Natural Lan-
guage Parsing: Psychological, Computational, and Theoretical Perspec-
tives (Dowty DR, Karttunen L, Zwicky AM, eds), 206–250. Cambridge:
Cambridge University Press.
Joshi AK, Schabes Y (1997) Tree adjoining grammars. In: Handbook of
Formal Languages, Vol. 3 (Rozenberg G, Salomaa A, eds), 69–126. Berlin:
Joshi AK, Vijay-ShankerK, Weir D (1991) The convergenceof mildly context-
sensitive grammar formalisms. In: Foundational Issues in Natural Lan-
guage Processing (Sells P, Shieber S, Wasow T, eds), 31–81. Cambridge,
MA: MIT Press.
Kauffman S (1993) The Origins of Order: Self-Organization and Selection in
Evolution. New York: Oxford University Press.
Kelso JAS (1995) Dynamic Patterns: The Self-Organization of Brain and
Behavior. Cambridge, MA: MIT Press.
Klingberg T (2009) The Overﬂowing Brain: Information Overload and the
Limits of Working Memory. Oxford: Oxford University Press.
Langer J (2000) The heterochronic evolution of primate cognitive develop-
ment. In: Biology, Brains, and Behavior: The Evolution of Human Devel-
opment (Parker ST, Langer J, McKinney ML, eds), 215–235. Santa Fe,
NM: School of American Research Press.
Lewontin RC (1998) The evolution of cognition: Questions we will never
answer. In: An Invitation to Cognitive Science. Vol. 4: Methods, Models,
and Conceptual Issues, 2nd ed. (Scarborough D, Sternberg S, eds), 107–
132. 2nd ed. Cambridge, MA: MIT Press.
Lieberman P (2006) Toward an Evolutionary Biology of Language. Cam-
bridge, MA: Harvard University Press.
Lorenzo G (2006) El vac´
ıo sexual, la tautolog´
ıa natural y la promesa mini-
malista. Madrid: Antonio Machado.
Marr D (1982) Vision. San Francisco: Freeman.
Minugh-Purvis N, McNamara KJ, eds (2001) Human Evolution Through
Developmental Change. Baltimore: Johns Hopkins University Press.
O’Donnell TJ, Hauser MD, Fitch WT (2005) Using mathematical mod-
els of language experimentally. Trends in Cognitive Sciences 9: 284–
Okanoya K (2007) Sexual display as a syntactic vehicle: The evolution of
syntax in birdsong and human language through sexual selection. In: The
Transition to Language (Wray A, ed), 46–62. Oxford: Oxford University
Oller DK (2005) The natural logic of communicative possibilities: Modularity
and presupposition. In: Modularity: Understanding the Development and
Evolution of Natural Complex Systems (Callebaut W, Rasskin-Gutman
D, eds), 409–434. Cambridge, MA: MIT Press.
Oster GF, Alberch P (1982) Evolution and bifurcation of developmental pro-
grams. Evolution 36: 444–459.
Owen R (1848) On the Archetype and Homologies of the Vertebrate Skeleton.
London: John Van Voorst.
Parker ST, Langer J, McKinney ML, eds (2000) Biology, Brains, and Be-
havior: The Evolution of Human Development. Santa Fe, NM: School of
American Research Press.
Parker ST, McKinney ML (1999) Origins of Intelligence: The Evolution of
Cognitive Developmentin Monkeys, Apes, and Humans. Baltimore: Johns
Hopkins University Press.
Perruchet P, Rey A (2005) Does the mastery of center-embedded linguistic
structures distinguish humans from non-human primates? Psychonomic
Bulletin and Review 12: 307–313.
Pinker S (1997) How the Mind Works. New York: Norton.
Ploog D (2002) Is the neural basis of vocalization different in non-human
primates and Homo sapiens? In: The Speciation of Homo Sapiens (Crow
TJ, ed), 121–135. London: British Academy.
Plotkin H (1997) Evolution in Mind: An Introduction to Evolutionary Psy-
chology. London: Alan Lane.
Pullum GK (1986) Footloose and context-free. Natural Language and Lin-
guistic Theory 4: 409–414.
Pullum GK, Rogers J (2006) Animal pattern-learning experiments:
Some mathematical background. http://www.lel.ed.ac.uk/∼gpullum/
Radzinski D (1991) Chinese number-names, tree adjoining languages, and
mild context-sensitivity. Computational Linguistics 17: 277–299.
Rasskin-Gutman D (2005) Modularity: Jumping forms within morphospace.
In: Modularity: Understanding the Development and Evolution of Natu-
ral Complex Systems (Callebaut W, Rasskin-Gutman D, eds), 207–219.
Cambridge, MA: MIT Press.
Reid RGB (2007) Biological Emergences: Evolution by Natural Experiment.
Cambridge, MA: MIT Press.
Reiner A, Brauth SE, Karten HJ (1984) Evolution of the amniote basal ganglia.
Trends in Neurosciences 7: 320–325.
Rochefort C, He X, Scotto-Lomassese S, Scharff C (2007) Recruitment of
FOXP2-expressing neurons to Area X varies during song development.
Developmental Neurobiology 67: 809–817.
Rogers J, Pullum GK (in press) Aural pattern recognition experiments and the
subregular hierarchy. UCLA Working Papers in Linguistics—Proceedings
of the Mathematics of Language 10.
Sasaki A, Sotnikova TD, Gainetdinov RR, Jarvis ED (2006) Social context-
dependent singing-regulated dopamine. Journal of Neuroscience 26:
Searle JR (1985) Minds, Brains and Science. Cambridge, MA: Harvard Uni-
Smith KK (2001) Heterochrony revisited: The evolution of developmental
sequences. Biological Journal of the Linnean Society 73: 169–186.
Smith KK (2002) Sequence heterochrony and the evolution of development.
Journal of Morphology 252: 82–97.
Striedter GF (2005) Principles of Brain Evolution. Sunderland, MA: Sinauer.
Striedter GF (2006) Pr´
ecis of Principles of Brain Evolution. Behavioral and
Brain Sciences 29: 1–36.
Thelen E, Smith LB (1994) A Dynamic Systems Approach to the Development
of Cognition and Action. Cambridge, MA: MIT Press.
Tettamanti M, Moro A, Messa C, Moresco RM, Rizzo G, Carpinelli A, Matar-
rese M, Fazio F, Perani D (2005) Basal ganglia and language: Phonology
modulates dopaminergic release. Brain Imaging 16: 397–401.
Todt D, Hultsch H (1998) How songbirds deal with large amounts of serial in-
formation: Retrieval rules suggest a hierarchical song memory. Biological
Cybernetics 79: 487–500.
Tomasello M, Farrar MJ (1986) Joint attention and early language. Child
Development 57: 1454–1463.
Uriagereka J (2008) Desperately evolving syntax. In: The Evolution of Lan-
guage: Proceedings of the 7th International Conference (EVOLANG7)
(Smith ADM, Smith K, Ferrer i Cancho R, eds), 331–337. Singapore:
Vijay-Shanker K, Weir D (1994) The equivalence of four extensions of
context-free grammars. Mathematical Systems Theory 27: 511–546.
Wagensberg J (2004) La rebeli ´
on de las formas o c´
omo perseverar cuando la
incertidumbre aprieta. Barcelona: Tusquets.
Weir D (1992) A geometric hierarchy beyond context-free languages. Theo-
retical Computer Science 104: 235–261.
Weir D (1994) Linear iterated pushdowns. Computational Intelligence 10:
304 Biological Theory 3(4) 2008