The Evolutionary Origin of The Language Areas in The Human Brain. A Neuroanatomical Perspective

.
Brain Research Reviews 25 1997 381396

Full-length review
The evolutionary origin of the language areas in the human brain. A
neuroanatomical perspective
Francisco Aboitiz
)
, Ricardo Garca V.
1
Programa de Morfologa, Instituto de Ciencias Biomedicas, Facultad de Medicina, Uniersidad de Chile. Independencia 1027, Casilla 70079, Santiago
07, Chile
Accepted 21 October 1997
Abstract
The capacity to learn syntactic rules is a hallmark of the human species, but whether this has been acquired by the process of natural
selection has been the subject of controversy. Furthermore, the cortical localization of linguistic capacities has prompted some authors to
suggest a modular representation of language in the brain. In this paper, we rather propose that the neural device involved in language is
embedded into a large-scale neurocognitive network comprising widespread connections between the temporal, parietal and frontal
. especially prefrontal cortices. This network is involved in the temporal organization of behavior and motor sequences, and in working
. active memory, a sort of short-term memory that participates in immediate cognitive processing. In human evolution, a precondition for
language was the establishment of strong corticocortical interactions in the postrolandic cortex that enabled the development of
. multimodal associations. Wernickes area originated as a converging place in which such associations concepts acquired a phonological
correlate. We postulate that these phonological representations projected into inferoparietal areas, which were connected to the incipient
Brocas area, thus forming a working memory circuit for processing and learning complex vocalizations. As a result of selective pressure
for learning capacity and memory storage, this device yielded a sophisticated system able to generate complicated utterances precursors
. of syntax as it became increasingly connected with other brain regions, especially in the prefrontal cortex. This view argues for a gradual
origin of the neural substrate for language as required by natural selection. q1997 Elsevier Science B.V.
Keywords: Brain evolution; Brocas area; Language; Parietal lobe; Prefrontal cortex; Temporal lobe; Syntax; Wernickes area; Working memory
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
2. Neural networks associated to language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
2.1. Brocas area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
2.2. Wernickes area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
2.3. The system for names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
2.4. Alleged connections between Wernickes and Brocas areas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
2.5. Monkey homologues to Brocas area and their connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
2.6. Homologues to Wernickes area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385
2.7. Some frontal connections of area Tpt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385
2.8. Evolutionary trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386
2.9. Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
3. Other frontal areas of interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
3.1. Prefrontal, granular frontal cortices and their temporoparietal afferents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
3.2. Working or active memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388
3.3. Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388
)
. Corresponding author. Fax: q56 2 678-6264 or 777-6916; E-mail: faboitiz@machi.med.uchile.cl
1
. . Tel.: q56 2 735-7068; fax: q56 2 678-6264 or 777-6916.
0165-0173r97r$32.00 q 1997 Elsevier Science B.V. All rights reserved.
. PII S0165- 0173 97 00053- 2
( ) F. Aboitiz, R. Garca V.rBrain Research Reiews 25 1997 381396 382
4. The evolutionary origin of language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389
4.1. Preconditions for language: the lexicon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389
4.2. The role of working memory in cross-modal associations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389
4.3. Origin of Wernickes area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390
4.4. The language rehearsal circuit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390
4.5. Granular frontal cortex: the central executive? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390
4.6. Acquisition of primitive syntactic rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
4.7. Neural representation of syntax: rules or learning abilities? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392
4.8. Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392
5. Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
5.1. The anatomy of language revisited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
5.2. Brain lateralization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
5.3. Final comment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
1. Introduction
The origin of human language has classically posed an
obstacle to the concept of human origins through natural or
sexual selection, originally because our communicative
capacities were considered to be related to spiritual matters
w x
74 . More recently the now widespread theory of universal
w x
grammar originally proposed by Chomsky 18 prescribes
that all languages, albeit superficially different, share com-
mon universal principles that are genetically determined
and unique with regard to other cognitive systems in the
brain. In other words, the rules that generate language are
considered to be fundamentally different from other cogni-
tive strategies that the brain may use. In this perspective
language is considered to be generated by a modular
language organ, that is informationally separated from
w x
other cognitive devices of the brain 18,31 . Considering
the evolutionary origins of language, Chomsky has argued
that such a system of rules is unlikely to have been
produced by a gradual process as required by natural
selection, since no hints of syntax precursors have been
observed in animals. Instead, language and the language
organ in the brain are proposed to have been originated as
a single biological macromutation. Nevertheless, a gradual-
w x
istic interpretation for the origin of syntactic rules 26,66
has emerged in the last years, in which the possibility of a
macromutation giving rise to all the complex set of gram-
matical rules is considered vanishingly small. These au-
thors agree with Chomsky in that syntactic rules do have a
genetic basis, but claim that they have been acquired as the
result of the gradual accumulation of small mutations.
Although the controversy on the evolutionary origin of
grammar is admittedly open, we will address certain issues
related to the neurobiological and neuroanatomical ele-
ments that make possible the generation of language. Our
perspective is that the neural device involved in language
is embedded in a large-scale corticocortical network link-
ing the parietotemporal higher-order sensory areas with the
frontal and prefrontal motor regions that participate in the
temporal ordering of behavior. Therefore, the concept of
language modularity must be considered with some cau-
tion as the language network is probably profusely inter-
connected with other, similar neural systems. Furthermore,
the concept of a language organ that has made a sudden
appearance in human evolution is difficult to reconcile
with the anatomical evidence. This rather suggests a grad-
ual specialization of the language-related regions from the
large-scale network mentioned above. Within this network
a sensory motor circuit processing linguistic stimuli and
controlling vocal output became locally differentiated, en-
abling primitive humans to generate increasingly complex
utterances. We will also argue that a sort of short-term
w x
memory defined as working memory 8 active memory
w x.
for others 34 has been fundamental in the evolution of
language.
In the paper, we will first describe the anatomy of the
language regions in the human brain and the connectivity
they may have considering the macaque brain as a model.
Then, the neural network for language will be proposed to
be included in the previously mentioned system of cor-
ticocortical connections that are essential for the organi-
zation of behavior. As said, one pivotal neurocognitive
process operating in this network is working memory. We
will suggest how the language regions may have differenti-
ated from a neural system specialized in processing com-
plex phonological sequences that were learned by imita-
tion. In doing so, the issues of the origin of the lexicon
from a primitive conceptual system that also requires
working memory, and of a primitive syntax will be ad-
dressed.
2. Neural networks associated to language
There are two relatively well-defined language areas
located in the left hemisphere of most right-handed people,
but perhaps more bilaterally in non-consistent right han-
w x
ders and left handers 20,86 . These are Brocas and
Wernickes areas, which very broadly speaking process
respectively motor and perceptual aspects of language.
Historically, these regions have been considered as some-
what isolated from other brain regions. Questions such as
whether there are other cognitive systems that relate to this
kind of network have been largely avoided. Perhaps, this
happened because of the technical difficulties in studying
the connectional anatomy of the human brain.
2.1. Brocas area
Brocas area is located in the pars triangularis and
opercularis of the inferior frontal gyrus Brodmanns areas

44 and 45; in the rest of the paper, all areas designated by
.
numbers correspond to Brodmanns nomenclature . True
Brocas aphasia consists of a non-fluent speech syndrome
in which agrammatism inability to handle and to generate

.
grammatical items is a key aspect. There are also difficul-
ties to understand grammatical elements such as reversible
passive sentences, and there is often distortion and mis-
.
placement of phonemes phonetic disintegration . Accord-
w x
ing to Damasio 20 this kind of aphasia occurs with
lesions in areas 44 and 45 and also prefrontal areas 6, 8, 9,
10 and 46 or some of them. Neurological damage re-
stricted to Brocas area produces a much more mild and
transient deficit in which the hallmark is a disturbance of
.
speech but not language aphemia . This may partly be a
consequence of the generalized observation that larger
lesions produce longer lasting and more profound effects
than smaller lesions because they assure more complete
damage or disconnection of areas 4445. However, the
evidence is also compatible with the concept that some
higher-level aspects of language structure require the in-
w x
tactness of frontal regions that surround Brocas area 20 ;
.
see Section 4.5 . Hence we will use the terms restricted
Brocas area when referring to areas 4445, and ex-
tended Brocas area when referring also to prefrontal
regions. In addition, it has been suggested that pari-
etofrontal networks as well as subcortical circuits includ-
ing the basal ganglia are of relevance for processing
w x
grammatical items 21 . This possibility is in accord with
connectional findings indicating a projection from the infe-
.
rior parietal lobe to Brocas region see Section 2.5 .
Brocas area is also a region for verb implementation
perhaps by virtue of its connections with prefrontal sectors
w x
involved in behavioral planning 20 .
2.2. Wernickes area
Wernickes area is in the posterior aspect of the supe-
rior temporal gyrus and in the floor of the posterior
Sylvian fissure the planum temporale, that corresponds to

the posterior moiety of Brodmanns area 22 or area Tpt as
w x.
defined by Galaburda and Sanides 35 . Lesions in Wer-
nickes area produce a fluent aphasic condition in which
the intonation and prosody of language are normal but
depending on the severity and location of the lesion, the
patients speech is difficult to comprehend because of
.
shifting of phonemes paraphasias and incapacity to find
the proper names of objects. Understanding sentences is
particularly difficult for most of these patients. However, it
is thought that more than a speech processing center,
Wernickes area is a place of convergence of cortico
cortical interactions in which sounds are mapped into
words and cross-modal associations occur with other sen-
w x
sory regions that convey meaning to these words 20 .
Language comprehension may occur in later stages of
cortical processing, perhaps involving wide areas of the
.
cerebral cortex including the parietal system . The regions
damaged in Wernickes aphasia are Brodmanns area 22
and usually the surrounding areas 37, 39 and 40.
w x
Early anatomical studies 38 indicated an asymmetry in
the human planum temporale, the planar region in the floor
of the Sylvian fissure behind Heschls gyrus that partly
corresponds to Wernickes area. The left side was larger
than the right in the majority of the cases. Furthermore,
angiographic and imaging studies have confirmed a rela-
tion between anatomical asymmetry in the posterior Syl-
vian fissure and functional lateralization as indexed by
w x
handedness or dichotic listening 2,3 . Recent gross
anatomical studies indicate that more than the quantitative
extension of the planum temporale, it is its arrangement in
relation to specific topographical landmarks that is asym-
w
metric, producing a qualitatively asymmetric image 2
x
4,48,50,53,84 . This situation may have biased the original
w x
study by Geschwind and Levitsky 38 , yielding a differ-
ence in planum size that perhaps more accurately can be
referred to as a difference in shape or position. These gross
morphological asymmetries apparently represent hemi-
spheric differences in the development of specific cortical
areas.
Recent cytoarchitectonic analyses of the superior tem-
poral lobe redefined the architectonic parcellation of this
w x
region 35,56,58 , describing three longitudinal streams or
lines of cytoarchitectonic differentiation along the superior
temporal lobe. There is a medial root line adjacent to the
insula in the depth of the Sylvian fissure, an intermediate
core line and a lateral belt line that runs in large part
.
along the superior temporal gyrus Fig. 1 . The latter
stream ends posteriorly in area Tpt, which is surrounded
by area PaAc at the end of the core line and by area
PaAe just a step behind Tpt in the same belt line in the
macaque, PaAlt corresponds topographically to human ar-
.
eas PaAe and PaAi . In humans, area Tpt is more extended
than in the macaque, occupies a large part of the planum
w x
temporale 35 and has been claimed to be highly asym-
w x
metric toward the left side 36 . Therefore given its
Fig. 1. Arrangement of architectonic areas in the superior temporal lobe
w x of the macaque. F.L., frontal lobe; T.P., temporal pole. Data from 33,56 .
anatomical location, it has been considered likely that area
Tpt corresponds in large part to Wernickes language
region in the left hemisphere other superior temporal

.
candidates are adjacent regions PaAc and PaAe .
2.3. The system for names
Naming deficits, particularly for proper nouns and emo-
tionally charged objects can occur with lesions of areas 21,
w x
20 and 38 of the anterior temporal lobe 20 . The inferior
temporal lobe is involved in processing the visual charac-
teristics of objects, and there are widespread connections
between different regions of the inferior temporal lobe and
the language areas of the superior temporal lobe which are
involved in the production of names for different objects
and attributes. Connections between anterior areas of the
inferior temporal lobethat are involved in mnemonic
and emotional aspects of visual processingand the pos-
terior language areas, appear involved in the generation of
w x
proper names and emotionally-charged objects 21,22 . On
the other hand, connections between more posterior as-
pects of the inferior temporal gyrus and Wernickes area
participate in the generation of category-specific names
w x
22 . For example, in humans a region in the lingual gyrus
.
of the inferior temporal lobe V4 is involved in color
processing, and disruption of the bilateral connection be-
tween this area and the left superior temporal lobe pro-
duces color anomia. In other words, a ring surrounding
the language areas in the lateral aspect of the temporal
lobe mediates the generation of names through cortico
.
cortical connections between temporal and parietal sen-
sory areas and the posterior part of the superior temporal
lobe. An additional region associated with Wernickes area
is the posterior insula, that also projects to the inferior
frontal gyrus and participates in auditory language process-
w x
ing 6,46 . A more recent report indicates that the left
precentral gyrus of the insula is involved in articulatory
w x
control for speech, which relates it to Brocas region 28 .
2.4. Alleged connections between Wernickes and Brocas
areas
Wernickes and Brocas areas are supposed to be inter-
connected through the arcuate fasciculus, a large tract of
long fibers that curves behind the bottom of the Sylvian
fissure and extends from the superior temporal region to
the inferior frontal gyrus however, connectional studies in

the monkey indicate a more complicated picture, see Sec-
.
tions 2.5 and 2.7 . Conduction aphasia involves lesions of
the arcuate fasciculus, of area 40 in the supramarginal
gyrus and even lesions of the insula that perhaps impair
fibers of the arcuate fasciculus in the underlying white
w x
matter 46 . This type of aphasia is characterized by the
incapacity to repeat sentences presented to the patient, a
characteristic that is shared by Brocas and Wernickes
aphasics but in the case of conduction aphasia there is
w x
preservation of speech production and comprehension 20 .
This is of interest in the context of this paper, because the
capacity to repeat sentences requires a sort of short-term
w x
memory 80 that in our view closely relates to working
memory. Furthermore, area 40 has been considered as a
temporary store of verbal information in working memory
w x .
46 ; see also Section 4.4 , and may have direct connec-
.
tions with Brocas region Section 2.8 . There are other
.
kinds of aphasia global, transcortical and other brain
regions involved in language processing thalamus, basal

ganglia, frontal mesial cortex, and even the non-dominant
. w x
hemisphere , but these will not be discussed here 19,20 .
2.5. Monkey homologues to Brocas area and their con-
nections
Since detailed connectional studies are impossible to
perform in humans, the most practical strategy has been to
use the macaque as a model of corticocortical connectiv-
ity of the human brain. Nevertheless, the macaque brain is
much simpler than the humans and there are difficulties
finding exact homologues of the human language areas in
.
the monkey see Fig. 2 . Of the areas composing Brocas
.
region Brodmanns areas 44 and 45 , only area 45 has
w x
been confirmed by Preuss and Goldman-Rakic 69 inside
the macaques inferior arcuate sulcus, between subareas
.
6V and 8Ar in prosimians neither area 44 or 45 exist .
w x
However, Barbas and Pandya 10 do not describe area 45
in the macaque, perhaps considering that the corresponding
region is part of premotor areas 6 or 8. In humans the
ventral part of premotor areas 6 and 8 is adjacent to areas
Fig. 2. Some architectonic areas of relevance to the present paper in the
. . human A and in the macaque B . Numbers correspond to Brodmanns
classification. AG, angular gyrus; AS, arcuate sulcus; CS, central sulcus;
IPS, intraparietal sulcus; PCS, precentral sulcus; PS, principal sulcus; SF,
sylvian fissure; SMG, supramarginal gyrus; STS, superior temporal sul-
cus.
4445, and therefore Brocas region can be conceived as a
differentiation of the ventral premotor region ventral area

.
6 of the monkey representing orofacial movements
w x
24,25,52,68 . Furthermore, it has been considered that
Brocas area is a premotor region specialized in the gener-
ation of complex rhythmic-hierarchic routines that gener-
w x
ate the motor coordination required for speech 39,45 .
Retrograde tracing studies indicate some temporal pro-
jections from the superior temporal gyrus and the superior
temporal sulcus, to the inferior post-arcuate and the pre-
w x
arcuate region 25 . A small part of this projection may
.
correspond to area Tpt Wernickes area . However, in
Section 2.7 it will be seen that anterograde tracing experi-
ments failed to detect a projection from area Tpt into the
inferior arcuate sulcus of the macaque. According to Preuss
w x
and Goldman-Rakic 71 , area 45 of the macaque receives
.
projections from parietal area 7b. This latter area 7b also
projects to the prefrontal cortex, in the intermediate princi-
. w x
pal sulcus of the frontal lobe area 46 62,71 . Another
interesting parietal area is 7ip within the intraparietal
sulcus, that projects to the dorsal and ventral aspects of the
.
anterior arcuate sulcus area 8 and to the posterior princi-
. w x
pal sulcus in the prefrontal cortex area 46 16,62,6971 .
.
The other components of parietal area 7 7m and 7a have
abundant connections with prefrontal cortex but a rela-
tively weak projection to the inferior bank of the arcuate
sulcus. In the inferior temporal lobe, area TE, involved in
object visual processing, also projects to the inferior arcu-
w x
ate sulcus, including area 45 11,13 . Although the role of
the latter in language processing is not clear yet, it may be
involved in working memory for object naming see Sec-

.
tion 4.5 .
2.6. Homologues to Wernickes area
.
We mentioned above Section 2.2 that one candidate
for Wernickes region is area Tpt in the superior temporal
.
lobe Fig. 2 . Surprisingly, area Tpt has been described in
w x
prosimians 70 , indicating that it may represent an ances-
.
tral character of primates more ancient than Brocas area .
Area Tpt of the macaque is a multimodal zone that has
w x
been claimed 57 to receive projections from auditory
association areas that surround it and from somatosensory
.
association areas in the parietal lobe but see Section 2.8 .
In the monkey, this area probably has a role in head-turn-
w x
ing movements oriented to localize sound sources 57,70 .
Visual association regions projecting to Tpt have not been
reported, but it is not clear if this has been extensively
investigated. The superior temporal lobe especially on the

.
left side is related to the processing of species-specific
w x
calls in the macaque 47,73 ; therefore it is possible that
area Tpt of the monkey is involved in the spatial localiza-
tion of conspecific calls.
2.7. Some frontal connections of area Tpt
In the macaque, area Tpt projects to the frontal lobe but
most of its terminals do not exactly reach the region that
w x
has been proposed to correspond to area 45 13,63 . Axons
from area Tpt end in dorsal area 8 dorsal prearcuate

.
region and in area 46 of the lateral frontal lobe the Tpt
moiety located in the surface of the superior temporal
.
gyrus , and in areas 8, 46 of the lateral frontal cortex and
also area 6 in the dorsal frontal lobe the Tpt moiety

located in the superior temporal plane, corresponding to
. w x w x
the planum temporale 63 . In the mentioned study 63 ,
none of three massive tracer injections in area Tpt were
able to produce labelling in the ventral arcuate sulcus
where the homologue of Brocas area is. Other, related
superior temporal areas such as PaAlt and ProA send their
projections to prefrontalor frontal granular areas 8, 9, 46
and 12. The only superior temporal region whose axons
were found in anterograde tracing studies to terminate in
the inferior arcuate sulcus corresponding to Brocas re-

.
gion were from ProA, that is the secondary auditory area
w x
63 .
w x
On the other hand, Deacon 25 indicates that a portion
of Tpt that extends into the superior temporal sulcus
. . Fig. 3. Projections of areas Tpt thick line , 7b dotted line and 7ip thin
. . line to the arcuate sulcus AS of the macaque. Abbreviations as in Fig.
w x 2. Data from 13,60,61 .
projects to the caudal bank of the inferior arcuate sulcus
.
possibly corresponding to premotor area 6 , interdigitating
.
with terminals from parietal regions area 7 . Deacon also
claims that regions in the anterior bank of the arcuate
.
sulcus corresponding to area 8 are predominantly associ-
ated with auditory association areas such as Tpt, while
.
postarcuate sites corresponding to area 6 tend to be
connected with somatic association areas like the inferior
parietal lobule. However, this interpretation is not consis-
tent with the higher resolution based on anterograde trac-

.
ing as opposed to retrograde tracing in Deacons work
studies mentioned in this section and in Section 2.5,
indicating that area Tpt and related ones are mainly con-
w x
nected with the dorsal moiety of the arcuate sulcus 63 ,
while parietal areas 7ip and 7b project to the anterior area

. .
8 and posterior area 6 banks of the inferior arcuate
w x
sulcus 16,62 . The overall evidence thus points to a much
more massive parietal than temporal projection to the
.
inferior arcuate sulcus in the macaque Fig. 3 . As will be
discussed in Section 4.3, this has important implications
for interpreting imaging studies of linguistic tasks per-
formed in humans. Nevertheless, anterograde tracing stud-
ies have the drawback that it is difficult to guarantee that a
particular area has been completely impregnated with the
tracer injection, while retrograde studies may better estab-
lish if a small region in a given area is sending projections
to an injected site. Therefore, the possibility of a small
subsector in area Tpt sending projections to the inferior
w x
arcuate sulcus 25 should not yet be dismissed.
2.8. Eolutionary trends
The subdivisions observed in area 7 of the monkey are
w x
absent in prosimians 70 . We suggest that the inferopari-
etal region has suffered a progressive differentiation in
primate phylogeny, and that areas 40 and 39 of the human
inferior parietal lobe arose as a consequence of further
subdivision of the inferior moiety of area 7 in the monkey.
If this is correct and if the connectivity in the monkey is
comparable to that of humans, areas 40 and 39or at least
part of themmay correspond to monkey areas 7b and 7ip
or related ones, and have important connections with
Brocas area. The latter makes functional sense as it
provides the basis for a somatosensory-motor circuit for
coordination of orofacial movements. On the other hand,
despite the gross-anatomical development of the temporal
lobe, higher-order areas in the superior temporal gyrus
such as Tpt and neighboring ones have been described as
somewhat conservative in all primates including humans
w x
35,71 . However, there is recent evidence indicating that
the cytoarchitectonic characteristics of area Tpt diverge in
humans with respect to other anthropoids in terms of
w x
increased spacing and number of neurons 14 .
w x
Pandya and Yeterian 57 indicate that in the macaque
area Tpt is connected with somatosensory parietal regions,
and a quantitative analysis of macaque corticocortical
w x
connectivity 85 places area Tpt as a node or bottleneck
connecting auditory regions with parietal areas including

.
7b and frontal regions. However, in a detailed and more
recent study of the monkey Cavada and Goldman-Rakic
w x
15 did not report inferoparietal projections to area Tpt,
indicating that perhaps in this species such connectivity is
not well developed. Considering that the vast majority of

corticocortical connections are reciprocal, a connection
from areas 7b or 7ip to area Tpt should be expected if the
.
latter projects to inferoparietal regions. There are abun-
dant projections from area 7ip to the posterior aspect of the
.
superior temporal sulcus superior and inferior banks and
some, very scarce, from areas 7b and 7a to the superior
w x
bank of the superior temporal sulcus 15 . Retrograde
w x
tracing studies 77 indicate some connections between
inferoparietal regions and the posterior most aspect of the
superior temporal sulcus, which may be adjacent to area
Tpt.
Overall, the above mentioned studies indicate that in the
macaque, the equivalent to Brocas area understood in a

.
restricted sense, only involving areas 4445 receives con-
nections from the inferoparietal lobe and inferior temporal
area TE, while the homologue of Wernickes area area

.
Tpt is rather connected to dorsal premotor regions and the
prefrontal cortex. We suggest that in human evolution, two
tendencies in the development of cortical connections oc-
curred that resulted in the origin of the language areas:
.
first and perhaps more important , the superior temporal
.
region area Tpt became increasingly connected with in-
feroparietal regions. This may have been achieved indi-
w x
rectly via regions in the superior temporal sulcus 15,77 or
the posterior insula which is interposed between temporal

w x.
and parietal cortices 5 , or through newly established
direct connections between area Tpt and the inferior pari-
etal lobe. This contributed to produce a strong link be-
tween the auditory system and a parieto-premotor loop
involved in the generation of complex vocalizations. The
development of inferoparietal sectors in the human brain
.
areas 3940 may partly relate to the establishment and
strengthening of this temporo-parietal link see Section

.
4.4 . The second process may have been the development
of direct connections between the precursor of Wernickes
region and areas 4445 and prefrontal sectors such as area
46, which served as a parallel pathway to send auditory
information to the orofacial premotor region.
2.9. Summary
Considering what is known about the connectivity of
the presumed equivalents to the language regions in the
monkey, the best candidates to Wernickes area seem not
to be massively connected with the candidates for Brocas
.
area inferior arcuate sulcus; Fig. 3 . Anatomical tracing
w x
studies 63 indicate that most of the frontal connections of
.
Tpt and related cortices representing Wernickes area are
with frontal granular cortex and with dorsal rather than
ventral area 8, which are not usually considered part of
.
Brocas language area in a restricted sense . In the
macaque, area 45 receives projections from parietal area
7b, and more generally the inferior parietal lobule projects
to cortical sectors related to ventral premotor areas 8 and 6
corresponding to Brocas area. Inferotemporal area TE
is also connected to area 45.
We therefore suggest that in human evolution, two sets
of projections developed from area Tpt andror neighbor-
ing regions: to the inferior parietal lobe and to prefrontal
and ventral premotor areas. This possibly established two
parallel connections between the incipient Brocas and
Wernickes areas: one direct and the other via the infer-
.
oparietal lobe Fig. 4 . In our view, the involvement of the
parietal lobe was a fundamental element in the evolution
of linguistic capacity, as it permitted the development of
an interphase between the auditory processing device and a
working memory circuit for complex vocalizations that
Fig. 4. Tentative network of connectivity for language in the human
brain. Especially important is the connection between SMG supramargi-
. nal gyrus and Brocas area, which may largely correspond to the arcuate
fasciculus. An additional area that may participate in language processing
is area TE, that projects to area 45. Connections between Tpt and Brocas
area, and between Tpt and SMG have not been substantially confirmed in
. the monkey but especially the latter are proposed to have developed in
the human lineage. Connections between area Tpt and SMG can be direct
. or indirect see text . FGC, frontal granular cortex. Other abbreviations as
in Fig. 2.
includes inferoparietal and ventral premotor regions see

.
Section 4.4 . On the other hand, a direct projection from
Wernickes region to frontal areas 4445 may have served
to transmit some automatic routines between Wernickes
and Brocas areas or to update the latter with auditory
information. Finally, direct projections from inferotempo-
w x
ral area TE to area 45 have been described 11 which may
serve as a third pathway that transmits linguistically rele-
w x
vant object information to Brocas area 39 .
3. Other frontal areas of interest
Although a common, very general plan of cortical
organization may exist in different mammals, there are
also important species and evolutionary differences in the
pattern of corticocortical connectivity and in the arrange-
ment of specific cortical areas. For example, a gross
morphological examination of a primate brain indicates
that higher-order cortical regions in the temporal, parietal
and especially in the frontal lobes have undergone a
notable increase in relation to other mammals this is

associated with a brain that is on the average about two
times larger than that of another mammal of the same size
w x.
1 . In the rest of the paper we will argue that some of
these expanding areas have direct relevance to the origin
of language.
3.1. Prefrontal, granular frontal cortices and their tem-
poroparietal afferents
Area 8 in the dorsal arcuate sulcus of the macaque is a
complex of premotor areas that includes the frontal eye
fields, involved in programmed eye movements. Since area
Tpt is considered to be involved in orienting head move-
ments, its connections to area 8 can be conceived as a
specialized sensorimotor loop involved in eye and head
movements that serves to orient attention toward sound
.
sources perhaps species-specific calls . Nevertheless, in
the context of this paper the most important frontal regions
are some areas collectively denominated as granular frontal
.
cortex including areas 9 and 46 . These participate in the
high-level organization of behavior and have undergone an
extensive growth in the anthropoid lineage no equivalents

to these areas have been reported in prosimians
w x.
33,34,68,69,71 .
However, more than the differentiation of specific corti-
cal regions, the development of long corticocortical con-
nections between the sensory association areas of the
parietal and temporal lobes and the frontal regions includ-
ing frontal granular cortex may better represent the modifi-
cations of the primate brain in regard to other mammals.
This is not to say that long, ipsilateral corticocortical
connections are not to be found in non-primates, but rather
that these connections are especially well developed in
primates. In the posterior cortex of primates and other
mammals modality-specific sensory areas are intercon-
nected in such a way that streams of information process-
ing can be described along each sensory modality
w x
51,56,58,60,87 . In addition, the different sensory areas
are connected with frontal regions via long, ipsilateral
axons. In the monkey, such temporoparietalprefrontal
w x
connections 58 are massive and organized topographi-
cally and reciprocally in such a way that the frontal cortex
has a map of the temporoparietal regions projecting to it,
and vice-versa. The anatomical arrangement of the lan-
guage areas fits this large-scale corticocortical network
and can be described as part of it. In this sense, the neural
architecture involved in language is embedded in a com-
plex system of large-scale connectivity that is the hallmark
of the primate brain, and therefore should not be consid-
ered as an isolated system working independently of simi-
larly organized corticocortical networks.
3.2. Working or actie memory
One of the functions of this system of long cortico
cortical connections is to integrate sensory and mnemonic
information of the temporoparietal lobes with the frontal
systems involved in the organization of behavior, both at
the short and the long-term. It may be described as a
conscious interphase between incoming sensory informa-
.
tion or recently retrieved information from memory stores
and the frontal system related to the execution of long-and
short term behavior. This sort of memory is currently
w x
referred to as working memory 8,4044,79 , and its
main feature is that the subject or the animal does not have
an immediate clue to solve the problem that is required as

.
in the case of classical associative memory but rather it
must keep in mind an object or event for some time
before executing whatever action is necessary for the
w x
memory task. Fuster 34 has denominated this system as
of active memory, indicating that it is a sort of short-term
memory that is used while performing specific behaviors.
Neurons in the prefrontal cortex that are specific for
certain stimuli become activated during the period in which
the monkey is rehearsing the short-term memory of an
event related to such stimuli. Some of these neurons tend
to extinguish their activity as the response time ap-
proaches, while others increase their firing as the interval
before the response comes to an end for review see Refs.

w x.
34,41 . This has been interpreted as a transition from
perceptually-related neuronal populations to behaviorally-
related populations in the frontal lobe. Furthermore, recent
w x
experiments 83 indicate that working memory for the
spatial location of objects involves the posterior parietal
.
cortex where spatial vision is processed and its connec-
tions with the dorsolateral prefrontal cortex, while working
memory for object characteristics relies on connections
between the inferior temporal lobe where object features

.
are processed and the inferior convexity of the prefrontal
cortex. Thus, working memory for different features is
segregated in different regions of the prefrontal and tem-
poroparietal lobes and their respective connections.
Working memory operates in concert with other cogni-
tive systems in the brain. It utilizes other types of memory
such as declarative or associative memory and holds these
in line while making decisions about appropriate responses
in specific environmental contexts. Furthermore,
w x
Goldman-Rakic 4042,44 claims that working memory is
the underlying neurobiological mechanism for most pre-
frontal functions, including tasks such as the Wisconsin
Card Sorting Test. This type of memory is essential for the
organization of behavior, as it guides it through representa-
tions of the outside world instead of relying directly on
external or sensory cues. For this reason, it is strongly
related to the formation of concepts and ideas and to our
capacity to think. The permanence of objects in our minds
relies in large part on working memory, which is funda-
mental for the formation of abstract concepts. In this sense,
working memory is a pivotal element in the establishment
of models or representations of reality.
3.3. Summary
From the preceding discussion, the language regions
and their connections appear embedded in a massive cor-
ticocortical network involved in higher brain functions
that appears to be the hallmark of the primate brain. One
main function performed by the temporoparietal networks
is working memory, which allows to keep on line several
cognitive items while performing specific behavioral tasks.
In Section 4 we will argue that the temporoparietalpre-
frontal networks together with corticocortical associa-
.
tions in the postrolandic cortex i contributed to establish
a conceptual system that provided an essential cognitive
.
precondition for the origin of language; and ii partici-
pated in the processing of increasingly complex linguistic
.
or prelinguistic utterances, thus contributing to the devel-
opment of the language areas and to the origin of a
primitive syntax.
4. The evolutionary origin of language
There are two especially striking features of language:
the ability to refer to objects or events in the external
world, and the rules that specify an order between the
different elements of a linguistic utterance. The former is
based on the lexicon, which prescribes distinct names for
different objects and events in the world; and the second is
the comprehensive set of grammatical rules or syntax any
w x
language has. Based on previous studies 37,82 and on the
evidence that apes trained in American Sign Language are
w x
capable to develop a lexicon but no syntax 67 , we will
argue that the origin of a lexicon required a specific neural
device that set the preconditions for language development
by the elaboration of a precursor of Wernickes area in the
superior temporal lobe. In addition, a circuit involving
.
mainly inferoparietalfrontal Brocas area projections and
connected to the incipient Wernickes region, developed as
a phonological rehearsal device that eventually differenti-
ated into the language areas and generated some basical
syntactic rules. Direct projections to Brocas area from
area TE and perhaps area Tpt may have also participated
in this process. Further development of temporoparietal
prefrontal circuits and the interactions between prefrontal
cortex and Brocas region permitted to achieve higher
syntactic levels and ultimately the generation of discourse.
4.1. Preconditions for language: the lexicon
w x
Geschwind 37 originally postulated a theory for the
origin of language that was based on the development of
non-limbic, cross-modal corticocortical interactions. Ac-
cording to him, these connections allowed to establish
associations between the sound of a vocalization and the
image of an object, permitting the generation of a lexicon
.
in which arbitrary sounds vocalizations represented ob-
jects identified through the visual or the tactile system.
Cross-modal auditoryvisual associations are perhaps the
best example for the origin of names, but Geschwinds
hypothesis was that in general, all sorts of cross-modal
.
and also unimodal associations were facilitated in the
human brain, allowing the origin of denotation through
vocal or gestural modes of communication. According to
Geschwinds hypothesis, in the monkey cross-modal sen-
sory associations need an intact limbic system to develop,
while in the human they become established via a short-cut
of corticocortical interactions. In fact, cross-modal asso-
ciations can be performed much faster in humans than in
w x
monkeys and apes 23 . Recent studies indicate that the
amygdala, a nuclear complex closely related to the limbic
system and located in the anterior temporal lobe, is essen-
tial for establishing cross-modal associations in the
macaque, while in the human lesions in the amygdala have
w x
much less effect in associative learning 5,54,55 .
Behavioral experiments performed during the seventies
and eighties showed clearly that if trained from an early
age, chimpanzees and gorillas are capable of learning some
aspects of sign language and can develop a reasonably
sized lexicon, although no hints of grammar could be
w x
observed 67 . This indicated that at least apes are able to
establish cross-modal associations such as those required
to generate a limited lexicon. Furthermore, there is evi-
dence that some primates such as vervet monkeys naturally
develop different cries specifying distinct predators, and
that the association between these cries and specific ani-
w x
mals is learned, not genetical 17,78 . Therefore, non-hu-
man primates have the potential to develop a naming
system under appropriate circumstances. It is likely then
that the neural substrate for the development for a lexicon
say, relatively well-developed cross-modal corticocorti-

.
cal interactions existed in an incipient form in higher
primates before the origin of humans, and was part of an
w x
adaptation for a different function 82 .
We suggest that in the case of higher primates and
hominids, natural selection directed evolution towards in-
creasingly developed corticocortical connections as a
preadaptation or in more modern terms, exaptation or

w x.
co-option 72 to the origin of a primitive lexicon. The
.
concept of preadaptation exaptationrco-option implies
that a structure or character that develops as a response to
a specific environmental condition may turn out to be of
utility in a different context. It is possible that in protohu-
mans, the development of complex cooperative and com-
petitive behavioral strategies between different members of
the social group put strong selective pressure for increased
brain plasticity and learning capacity, which included the
development of corticocortical connections. Furthermore,
if vocalizations were useful to coordinate behaviors during
food gathering, sexual relations or in any other way,
cross-modal corticocortical associations between the vi-
sual and the auditory systems were not a difficult step to
achieve and the first names were originated. From then on,
selective pressure concentrated on the development of
specific types of corticocortical connections that permit-
ted the establishment of cross-modal associations and al-
lowed to generate a lexicon. Another but not necessarily
alternative possibility is that the development of manual
skills for fabricating tools enhanced different types of
cross-modal associations visualsomatosensory, visual

.
auditory and somatosensoryauditory , which enabled the
possibility to generate integrated concepts involving sev-
w x
eral modalities 82 .
4.2. The role of working memory in cross-modal associa-
tions
Cross-modal associations were important also to pro-
duce internal, multimodal representations of objects and
.
places concepts , essential for a complex social life
w x
21,49,82 . Multimodal concepts may facilitate the retrieval
of names from an internal representation rather than from
direct sensory input, thus enhancing working memory
capacity. On the other hand, working memory powered the
ability to form cross-modal associations since it allows
information coming in at one point in time to be tran-
siently associated with information encountered at a differ-
w x
ent time 40,41 . As a consequence, the timewindow for
the establishment of associative memories is expanded,
therefore enhancing the probability to form incipient links
between different stimuli and contributing to establish
multimodal representations. Furthermore, working memory
was fundamental for the acquisition of a reasonably sized
lexicon, as it allowed to hold concepts on line while
performing linguistic behavior, thus facilitating say, audi-
toryvisual associations. In this way, the formation of
w x
concepts 21 and a primitive lexicon depended not only on
cross-modal associations but also on a powerful working
memory system and on the interaction between these two
memory devices.
4.3. Origin of Wernickes area
The region corresponding to the parietooccipitaltem-
.
poral junction roughly corresponding to Wernickes area
has been proposed as a place of convergence for cortico
cortical interactions, such that not only associations be-
tween different modalities take place but also associations
w x
between associations 37,82 . We postulate here that espe-
cially the superior temporal region differentiated as a
bottleneck for cross-modal associations that in this place
acquired a phonological correlate. In this way, Wernickes
area originated as a place in which multimodal representa-
tions or concepts obtained a linguistic dimension by being
w x
mapped into simple phonological sequences 20 .
4.4. The language rehearsal circuit
We have discussed the origin of the lexicon through an
efficient system of cross-modal interactions and working
memory, and the development of Wernickes area as a
point of convergence of multimodal associations. We will
now suggest that Brocas area and its inferoparietal coun-
.
terpart see Sections 2.5 and 2.8 originally developed as
part of a neural circuit involved in performing complex
vocalizations. In order to analyze this subject in proper
detail, it will be useful first to discuss some recent results
on brain activation during linguistic working memory tasks.
Working memory has been proposed to consist of two
main components: slave systems involving working mem-
ory of the specific sensorimotor circuits involved in partic-
ular behaviors, and a generalized, multimodal central exec-
utive that controls and distributes attention according to
w x
the contingencies 8,9 . A slightly different view is that of
w x
Goldman-Rakic 43 , who asserts that instead of a pan-
modal central executive controlling several modality-
specific slave systems, there are multiple special-purpose
working memory systems organized in parallel. This possi-
bility will be discussed in Section 4.5.
In the human, some imaging studies have been per-
formed in order to determine the localization of the so-
called central executive and slave systems, especially for
w x
verbal working memory 7,32,46 . Using this methodol-
ogy, a slave linguistic phonologicalrehearsal working
memory system has been allocated to the supramarginal
.
gyrus parietal area 40 and Brocas areathe former
more involved in the phonological store and the latter in
the subvocal rehearsal that takes place during verbal work-
w x w x
ing memory 59,75 . Paulesu et al. 59 used two different
tasks, one requiring short-term memorization of letters
involving both a phonological store and a subvocal re-

.
hearsing system and the other requiring rhyming judge-
ment for letters involving only the subvocal rehearsal

.
system . These tasks were compared to non-verbal work-
ing memory reference tasks. In this experiment, both the
control and the experimental tasks involved working mem-
ory, while only the experimental situation required linguis-
tic processing. The results indicated activation of the
.
supramarginal gyrus in the parietal lobe area 40 in
relation to the phonological store, and activation of Brocas
area in relation to the subvocal rehearsal. It is of the
highest interest that the regions involved in this phono-
logical loop closely correspond to the connections above
proposed between the inferior parietal lobe and Brocas
.
area Sections 2.5 and 2.8 . Furthermore, we discussed that
lesions in area 40 and the arcuate fascicules presumably

.
connecting the inferior parietal lobe with Brocas area
w x
produce conduction aphasia 20 , which may be considered
as a deficit in verbal working memory.
From the studies in the monkey reviewed before Sec-

.
tions 2.7 and 2.8 , frontal projections from area Tpt
.
Wernickes region do not terminate massively in Brocas
.
area in the restricted sense of areas 4445 . In Section 2.8
we initially proposed that in human evolution area Tpt
may have become increasingly connected with inferopari-
.
etal regions such as the supramarginal gyrus area 40 thus
feeding the latter with auditory information to be used in
.
the phonological loop see Fig. 4 . It is therefore possible
that in the human the posterior superior temporal region
represents a transitional zone in which concepts progres-
.
sively acquire a phonological correlate see Section 4.3 ,
while the supramarginal gyrus area 40, and perhaps also

.
the angular gyrus, area 39 in the parietal lobe stores this
phonological representation for a brief time. Some neurons
.
in human area 40 and perhaps in area 39 may then
.
project to Brocas region areas 44 and 45 , thus establish-
ing a neuronal circuit for the phonologicalrehearsal sys-
tem of linguistic working memory. If they exist, direct
projections from Tpt or neighboring areas to Brocas re-
gion may participate in language processing in at least two
.
possible ways: i generating a shortcut between Wernickes
.
and Brocas regions for some automatic routines, and ii
participating in higher levels of language processing see

.
Section 4.5 .
4.5. Granular frontal cortex: the central executie?
On the other hand, granular frontal areas area 9 and

.
especially area 46 relate with more general aspects of
working memory, perhaps associated with the central exec-
w x
utive 61,64,65,75 . For example, in the study by Petrides
w x
et al. 65 , the experimental tasks involved verbal working
memory processes i.e., saying aloud the numbers 1 to 10

.
but in random order and without repeating each other
while the control task consisted of just counting from 1 to
10, which minimizes the component of working memory
as it is an automatic procedure. In this experiment, both
the experimental and the control tasks were linguistic,
while working memory was required only for the experi-
mental task. Therefore, in the latter only the central execu-
tive was expected to be activated since this should be
specific for working memory tasks and not depend on
particular sensorimotor circuits. The result was an in-
creased blood flow bilaterally in the mid dorsolateral
frontal cortex, which corresponds to the anatomical loca-
tion of areas 9 and 46.
There is however some controversy as to the existence
of an anatomically delimited central executive in the pre-
w x
frontal cortex 42,43,61,81 . As mentioned, Goldman-Rakic
w x
43 claims that the prefrontal cortex is subdivided into a
mosaic of areas that are specialized in working memory
for different modalities, there being no specific anatomical
location for a multimodal central executive that coordi-
nates attention. Thus, visuospatial working memory in-
volves connections between parietal area 7 and prefrontal
areas 46 and 8; while working memory for object features
depends on inferotemporal area TE and its connections
w x
with frontal areas 45 and 12 83 ; finally, linguistic work-
ing memory would involve connections between infer-
oparietal areas 3940 and frontal areas 4447 which

roughly correspond to the phonologicalrehearsal loop
.
mentioned above . Based on the data from brain lesions
and the connectional information discussed above Sec-

.
tions 2.1, 2.5 and 2.7 , we propose that beside areas 44r45
.
and 47 , frontal granular areas such as 9 and 46 also
participate in language processing, especially in aspects
related to working memory tasks this fits the notion of an

.
extended Brocas area . Perhaps these frontal granular
areas not only relate to the distribution of attention but also
.
handle cognitive semantic information that is relevant for
language processing. For example, when recalling the ob-
jects observed in a room, one might say there is a lamp
.
with a red screen object feature information in the left
.
bottom corner visuospatial information . Tasks such as
this probably require the coordinated activity of the respec-
tive working memory circuits that are located in granular
frontal cortex. In this sense, although the so-called central
executive is in fact a mosaic of different regions subserv-
ing different modalities, these areas probably cooperate
with the circuit for linguistic processing by feeding cogni-
.
tive information and distributing the activity attention
between the different modalities. Furthermore, there are
direct projections from areas in the superior temporal lobe
.
including area Tpt, Section 2.7 and the inferior parietal
.
lobe to granular frontal cortex see Fig. 4 , indicating that
a degree of topographic overlap of the different working
memory systems exists in the prefrontal cortex although

perhaps there is a microscopic segregation of the respec-
.
tive terminals . In this context, the central executive may
be conceived as a mosaic of closely interconnected but
functionally diverse prefrontal regions, whose activities
become tightly coordinated during complex working mem-
ory tasks. Therefore, we should expect activation of sev-
eral frontal granular areas in many types of relatively
complicated working memory paradigms, especially for
the case of linguistic working memory as it requires
.
handling several and varied items of cognitive semantic
w x
information 65 . In the origin of language, the capacity for
conscious handling and coordination of spatial, object
feature and linguistic information in a short period of time
may have been a highly adaptive character as it enor-
mously facilitated communication.
This whole consideration implies that language process-
.
ing is closely linked to working memory i in terms of the
anatomical arrangement of the neural networks involved,
.
and ii because it operates in the context of an efficient
working memory system. One of our main suggestions is
that selective pressure for the capacity to learn complex
vocalizations through imitation and repeated practice was a
.
key aspect in establishing a slave phonological working
memory system that allowed to temporarily store phono-
logical representations in order to rehearse them internally.
Through the action of natural selection favouring good
learners, this system eventually differentiated into some
primordial language regions. Concomitantly, a prefrontal
system in which information from other sensory modalities
was integrated and coordinated with the representation of
complex vocalizations was also being developed see Sec-

.
tion 4.6 .
4.6. Acquisition of primitie syntactic rules
As mentioned in Section 2.5, it has been proposed that
the elementary aspects of syntax arose as a consequence of
the differentiation of the premotor regions related to
w x
Brocas area 24,25,39,45,52,68 , in which these regions
specialized in controlling complex, high-frequency se-
quences of movements such as those involved in articulate
speech. This fits the concept of temporoparietalprefrontal
networks involved in the temporal organization of behav-
w x
ior and motor patterns 33 . Similar to other accurate
sensorimotor systems where sensory feedback is of great
relevance for the performance of complex, learned tasks
w x
27 , projections between auditory association areas like
. .
area Tpt and parietal somatosensory areas like area 40 ,
and from these regions to the premotor orofacial cortex
.
Brocas area were important for learning socially rele-
vant vocalizations. Supporting our view, Damasio and
w x
Damasio 21 suggest that the neural representation of
grammar is related to Brocas area but also involves
regions related to the parietal lobes and the basal ganglia,
the latter involved in the automatization of certain motor
routines.
The above paragraph does not imply, however, that
syntax is solely represented in the slave system subserv-
ing the above mentioned phonological loop. Recall that for
true Brocas aphasia to occur damage in the frontal granu-
w x
lar cortex is required 20 , which indicates that grammar is
not only located in the classically-defined language areas
. w x
Section 2.1 . Fuster 33 and others have proposed that in
the cerebral cortex, the sensory and motor processing
hierarchies are interconnected at different levels, thus gen-
erating distinct domains of sensorimotor integration. It is
well possible that in a broad sense, a complex syntax
results from the coordinated operation of these different
hierarchical domains, from the lowest phonememorpheme
levels to higher levels involving sentence syntax and even-
tually the organization of discourse. In this way, although
some basic components of syntactic rules including

.
phonememorpheme relations are directly related to the
language regionsthe phonological rehearsal loop, the
interactions between these areas and brain components
involved in other cognitive tasks especially granular frontal

.
cortex is probably fundamental in the development of
higher-level domains of grammar. The distribution and
coordination of activities in the different prefrontal devices
.
granular frontal cortex and Brocas area probably im-
poses processing constraints that result in superior levels
of syntax. In our view, semantic properties such as causal-
ity, object permanence, time order and others appear in the
syntactic space through integration of the respective cogni-
tive systems in the temporoparietalprefrontal networks
.
and the phonological loop Section 4.3 . This indicates that
multimodal concepts that conform the semantic space in
the temporoparietal lobes reach on one hand Wernickes
area, then the inferoparietal cortex and finally Brocas
area; and on the other hand they are shuttled directly to the
prefrontal cortex in order to interact with the rehearsal
system in Brocas area, generating higher-level syntactic
rules.
4.7. Neural representation of syntax: rules or learning
abilities?
Initially, the advantage of a phonological sensorimotor
circuit was to permit the rehearsal and the temporary
storage of increasingly complex vocal utterances, involv-
ing activity in Brocas area and its parietal counterpart.
Although in ontogeny this originally appeared as transient
.
working memory, its continued and eventually automatic
use ended stabilizing certain neural networks involving
Brocas and Wernickes areas and parietal regions plus

.
other, subcortical regions , that resulted in the acquisition
of an internal representation of complex phonological se-
.
quences a primitive grammar . This may have occurred
partly as a process of neuronal group selection in which
specific neural networks become differentially stabilized
according to their capacity to generate recurrent activation
w x
29 . Modern connectionist models based on networks that
learn to recognize patterns on the basis of changing synap-
tic weights are consistent with a group selective process
and have been recently proposed to account for several
w x
aspects of language learning 76 . These kinds of models
perhaps represent the initial process of language acquisi-
tion in primitive hominids as well as in modern humans.
In phylogeny however, the situation is slightly different
as this requires heritable, genetic changes. Not all individ-
uals were alike in their ability to learn complex protolin-
guistic utterances. Those who better developed the appro-
priate neural networks were more able to learn and under-
stand such utterances, which gave them an advantage over
others. Part of this variability in learning capacity was
genetic, and therefore the result was an evolutionary trend
to increase the capacity to learn syntactic rules. If this
proposal is correct, language learning is genetically deter-
mined but on the other hand syntax especially high-level

.
syntax does not come imprinted in the genetic system; it
must rather be learned through social exposure. In other
words, we have become just very good to learn grammati-
.
cal rules genetically inclined to do so , but we do not bear
a genetic representation of such rules we are not born

.
with an internal representation of grammar . If truly uni-
versal and specific for language, these rules represent
learning constraints determined by the peculiarities of the
dynamical working-memory circuits involved in language
acquisition, or constraints determined by the integration
between the purely vocalphonological system and more
widespread neurocognitive networks that we postulate gen-
w x
erate higher levels of syntax 76 .
In some cases, natural selection favours the develop-
ment of specific, genetically-determined neuronal circuits
w x
1 . Why cant this happen with circuits encoding gram-
matical representations? It has been proposed that such
rules can be canalized into the genome by the action of
w x
natural selection 66 . One caveat regarding this assump-
tion is whether there has been sufficient time for natural
selection to develop this canalization process. More impor-
tant, although we are not claiming that the incorporation of
syntactic rules in the genome is impossible, we consider
that in the case of language acquisition the learning com-
ponent and the exposure to speakers are so important that
perhaps selection has been in the direction of increasing
learning capacity rather than specifying neuronal circuits
that develop autonomously. Although it has been claimed
that children have an innate knowledge of syntax, evidence
is by no means clear as to what extent they are capable of
learning grammar without proper exposure to language. In
other words, in order to substantiate the claim that there is
a genetic representation of grammar there should be strong
evidence indicating that linguistic performance can de-
velop autonomously, evidence that is nonetheless elusive
w x
76 .
4.8. Summary
The capacities for cross-modal associations and work-
ing memory probably preexisted in primates before the
origin of language this explains the linguistic abilities of

some apes and the diversity of warning calls of vervet
.
monkeys , but in the hominid lineage they developed much
further. This permitted the origin of a sophisticated con-
ceptual system that included an elaborated, internal repre-
sentation of the world, which allowed the emergence of
primitive linguistic elements denotating particular objects,
actions or places. Wernickes region appeared as a place of
convergence for multimodal concepts, where the latter
acquired a phonological dimension.
In a relatively complex social system such as that of
primitive hominids, probably there was a diversity of
learned vocalizations and calls that had to be learned by
imitation by youngsters. The projections comprising the
phonological rehearsal loop between the inferior parietal

.
lobe and the precursor of Brocas area originated as a
working memory device involved in the imitation of differ-
ent vocalizations. This system must have been fed by
auditory information, either via superotemporal projections
to the parietal lobe, andror via the scant auditory termina-
tions in the ventral arcuate sulcus.
Initially, the incipient language areas were able to coor-
dinate phonemes and morphemes into primitive words, and
eventually some sequences of words. Higher levels of
syntax, such as those required to formulate more complex
sentences and ultimately to produce discourse were partly
consequence of the differentiation of the anterior language
areas into a timing device generating complex, hierarchi-
w x
cally organized motor sequences 45 , but also resulted
from the integration of the language areas with other
cognitive systems in the brain. Projections from the pari-
etal and the temporal lobes into frontal granular areas were
probably very important in the achievement of these higher
syntactic levels. The prefrontal system thus allowed to
maintain into attention the semantic contents of complex
linguistic utterances while they were being generated, and
set constraints for the development of higher syntactic
rules. In the individual, complex sequences of vocaliza-
tions originally appeared as transient memories that be-
came eventually stabilized as long-term representations.
Natural selection enhanced the capacity to learn such
sequences and consequently grammar became increasingly
complex.
.
In synthesis, i the development of cross-modal associ-
ations and the conceptual system in the postrolandic cortex
provided a basis for a semantic neural device, that con-
.
verged into a region prospective Wernickes area in
which these concepts acquired a phonological correlate.
.
This phonological representation was ii transferred into a
phonologicalrehearsal apparatus that included the inferior
parietal lobe and the premotor representation of orofacial
movements, providing some basic syntactic rules at the
levels of phonemes and morphemes. It is also possible that
direct projections from area Tpt to areas 4445 also
.
developed, providing a parallel circuit. Finally, iii the
coordinated operation of networks involving granular
frontal cortex and the semantic system represented in the
temporoparietal lobes, together with the phonologicalre-
hearsal loop just mentioned, generated higher levels of
syntax and discourse.
5. Discussion
In this paper we have outlined a hypothesis for the
evolutionary origin of language starting from a large-scale
corticocortical network that connects different sensory
regions among themselves as well as sensory regions of
postrolandic cortex with frontal regions that participate in
the organization of behavior. Therefore, we conclude that
the neural substrate for language may well have arisen
through the gradual differentiation of a preexisting system
in the primate brain. This device by no means can be
considered informationally separated from other cognitive
systems, although it is possible that there are certain neural
processes that are specific for the regions involved in
language processing.
Although we are proposing a certain sequence in which
the capacity to give names and the development of Wer-
nickes area as a multimodal zone preceded the origin of
syntax and the differentiation of Brocas area, while higher
levels of syntax involving granular frontal cortex were the
last step to achieve, there was probably a significant
overlap in the different steps. Furthermore, we cannot state
with certainty when in human evolution this happened.
The arguments for or against the existence of language in
w x
different hominids 30 are still open to many interpreta-
tions. A critical aspect regarding the origin of syntax is
that in our view it appeared in our ancestors as transient
sensorimotor coordination that was initially acquired by
imitation. After continued use, in each individual this
transient neural circuit became stabilized in an internal
representation. On the other hand, in the phylogenetic
history of humans there was selection for increased learn-
ing capacity rather than for the genetic specification of a
set of linguistic rules. Therefore, the genetic substrate for
language can better be conceived of as a learning strategy
rather than as a set of genetically-determined syntactic
structures. A consequence of this situation is that the same
learning device that was present in proto-language acquisi-
tion in early humans is still at work, only much more
elaborated.
5.1. The anatomy of language reisited
Although it is clear that several other systems contribute
w x
importantly to communication and language 19,20 , we
have emphasized the regions corresponding to Brocas and
Wernickes areas because these are the sectors most di-
rectly related with the paradigmatic aspects of language
that are the lexicon and syntax. One important aspect of
the connectivity of these regions is that in the macaque the
.
equivalent of Brocas area area 45 receives major projec-
tions from the inferior parietal and the inferior temporal
lobes instead of the superior temporal lobe, where the
equivalent of Wernickes region is supposed to be. The
inferoparietal areas from which some of these projections
.
arise in the monkey areas 7b and 7ip are proposed here
.
to be homologous to areas 40 supramarginal gyrus and
.
perhaps 39 angular gyrus in the human. We propose that
.
in humans, Wernickes area Tpt and neighboring ones
.
feeds directly or indirectly areas 40, 39 or some infer-
oparietal regions that project to Brocas area. It also pro-
jects directly to prefrontal cortex Brocas area in an

.
extended version , thus participating in the central execu-
tive for language working memory. Direct projections
.
from Tpt to areas 4445 restricted Brocas area are
scarce if they exist in the macaque. These connections may
have become strengthened in the human lineage, perhaps
participating in the generation of some automatic linguistic
sequences. Furthermore, direct connections between infer-
otemporal area TE and Brocas region may be a third
pathway to transmit linguistically relevant object informa-
tion. Therefore, the connectivity of the language regions
may be more complicated than the currently accepted
model of Brocas and Wernickes areas directly connected
via the arcuate fascicules. Brocas region appears to be, at
least cytoarchitectonically, a recently originated region in
the history of primates something similar occurs with

.
areas 40 and 39 . On the other hand, the region corre-
sponding to Wernickes area has been described as an
w x
ancient structure in primates 69 , although as said there is
evidence indicating cytoarchitectonic differentiation of area
w x
Tpt in the human with respect to other primates 14 .
These considerations about connectivity have the draw-
back of being based on a model animal that does not
speak, and therefore they may not correspond precisely to
the connectivity in the human language areas. Further-
more, further anatomical and functional studies may deter-
mine that the language circuits include some different
cortical regions and connections than those specified at
this moment. Ours is intended to be a starting model in
.
which the connectivity of the language regions i is
considered to be more complex than only a direct connec-
tion from Wernickes region in the temporal lobe to Brocas
area through the arcuate fascicules, as it includes at least
.
the inferoparietal system; and ii is tightly related to
circuits for working memory in the prefrontal cortex.
5.2. Brain lateralization
Another critical issue in the evolution of the language
regions is that they tend to be localized in the left hemi-
sphere. The explanation for this asymmetry is beyond the
scope of this paper, but it may be a consequence of
previously established asymmetries for sequential left

. .
hemisphere vs. spatial right hemisphere skills. Perhaps
the manufacturing of tools imposed a left-hemisphere ten-
dency for sequential movements, to which the language
w x
system turned out to fit 12 . Whatever the reason for
language lateralization, we propose that it mainly consists
of the specialization of two different types of temporopari-
etalprefrontal projections: the left hemisphere emphasizes
temporal and inferoparietal connections with the frontal
.
lobe sequentialr linguistic processing , while the right
emphasizes projections from posterior parietal areas in-
volved in spatial vision. This perhaps explains the rela-
tively modest differences in anatomical arrangement be-
tween the language regions in the left hemisphere and their
w x
right-hemisphere equivalents 2 , as compared to the func-
tional laterality of the two hemispheres. In other words, the
structural basis for hemispheric asymmetry may lie in the
arrangement of temporoparietalfrontal projections in the
two hemispheres rather than on the development of spe-
cific cortical areas.
5.3. Final comment
Finally, and putting this contribution in a broader con-
text, our perspective permits to understand the origin of
language as a gradual process since the underlying neural
structure for language is present in non-human primates.
The development of natural languages would result from a
refinement of connectivity patterns already incipient in
other primates that became organized according to selec-
tive pressure on learning capacity, and that generated
syntax on the basis of increasing complexity of the respec-
tive cortical circuits. In this sense, language gradually
originates from a series of preexistent devices that in
retrospect can be seen as preadaptations for linguistic
capacity. Making emphasis on the development of neu-
roanatomical networks involved in higher-level processing
rather than on the origin of syntactic rules, this view is
consistent with a darwinian perspective on the origin of
language rather than with the chomskian concept of a
discrete language organ that has arisen by a single macro-
mutational event.
Acknowledgements
We are indebted with Cristian Warnken from ARTV
Chile, as an interview with him partly motivated us to
write this article. In addition, we are especially grateful to
Joaqun Fuster for his thoughtful comments. Guillermo
Soto and Ximena Carrasco also provided useful sugges-
tions during the preparation of the manuscript. Finally,
anonymous reviewers provided constructive criticism and
suggestions on the views presented here. Artwork was
made by Samuel Valenzuela excepting Fig. 4 that was
prepared by Claudia Andrade. Supported by FONDECYT
.
grant No 1970294 Chile .
References
w x 1 F. Aboitiz, Does bigger mean better? Evolutionary determinants of
. brain size and structure, Brain Behav. Evol. 47 1996 225245.
w x 2 F. Aboitiz, A. Ide, Anatomical asymmetries in language-related
cortex and their relation to callosal function, in: H. Whitaker, B.
. Stemmer Eds. , Handbook of Neurolinguistics, Academic Press,
New York, in press.
w x 3 F. Aboitiz, A.B. Scheibel, E. Zaidel, Morphometry of the Sylvian
fissure and the human corpus callosum, with emphasis on sex
. differences, Brain 115 1992 15211541.
w x 4 F. Aboitiz, A. Ide, R. Olivares, Corpus callosum morphology in
relation to cerebral asymmetries in the post mortem human, in: E.
. Zaidel, A. Pascual-Leone, M. Iacoboni Eds. , Structure and Func-
tion of the Human Corpus Callosum: a Case Study in Cognitive
Neuroscience, Academic Press, New York, in press.
w x 5 J.P. Aggleton, The contribution of the amygdala to normal and
. abnormal emotional states, Trends Neurosci. 16 1993 328334.
w x 6 J.R. Augustine, Circuitry and functional aspects of the insular lobe
. in primates including humans, Brain Res. Rev. 22 1996 229244.
w x 7 E. Awh, E.E. Smith, J. Jonides, Human rehearsal processes and the
. frontal lobes: PET evidence, Ann. New York Acad. Sci. 769 1995
97118.
w x 8 A.D. Baddeley, Working Memory, Clarendon Press, Oxford, 1986,
289 pp.
w x 9 A.D. Baddeley, S. Della Sala, Working memory and executive
. . control, Phil. Trans. Roy. Soc. London 351 1996 13971402.
w x 10 H. Barbas, D.N. Pandya, Architecture and intrinsic connections of
the prefrontal cortex in the rhesus monkey, J. Comp. Neurol. 286
. 1989 353375.
w x 11 J.F. Bates, F.A.W. Wilson, S.P. O Scalaidhe, P.S. Goldman-Rakic,
Area TE connections with inferior prefrontal regions responsive to
. complex objects and faces, Soc. Neurosci. Abstr. 20 1994 434.10.
w x 12 J.L. Bradshaw, L.J. Rogers, The Evolution of Lateral Asymmetries,
Language, Tool Use and Intellect, Academic Press, San Diego,
1992, 463 pp.
w x 13 J. Bullier, J.D. Schall, A. Morel, Functional streams in occipito
. frontal connections in the monkey, Behav. Brain Res. 76 1996
8997.
w x 14 D. Buxhoeveden, W. Lefkowitz, P. Loats, E. Armstrong, The linear
organization of cell columns in human and nonhuman anthropoid
. Tpt cortex, Anat. Embryol. 194 1996 2336.
w x 15 C. Cavada, P. Goldman-Rakic, Posterior parietal cortex in rhesus
monkey: I. Parcellation of areas based on distinctive limbic and
. sensory corticocortical connections, J. Comp. Neurol. 287 1989
393421.
w x 16 C. Cavada, P. Goldman-Rakic, Posterior parietal cortex in rhesus
monkey: II. Evidence for segregated corticocortical networks link-
ing sensory and limbic areas with the frontal lobe, J. Comp. Neurol.
. 287 1989 422445.
w x 17 L. Cheney, R.M. Seyfarth, How Monkeys See the World, Univ. of
Chicago Press, Chicago, 1990, 377 pp.
w x 18 N. Chomsky, Rules and Representations, Columbia Univ. Press,
New York, 1978, 299 pp.
w x 19 B. Crosson, Subcortical Functions in Language and Memory, Guil-
ford Press, New York, 1992, 374 pp.
w x . 20 A.R. Damasio, Aphasia, N. Engl. J. Med. 326 1992 531539.
w x 21 A.R. Damasio, H. Damasio, Brain and language, Sci. Am. 267
. 1992 8895.
w x 22 H. Damasio, T.J. Grabowski, D. Tranel, R.D. Hichwa, A.R. Dama-
. sio, A neural basis for lexical retreival, Nature 380 1996 499505.
w x 23 R.K. Davenport, Cross-modal perception in apes, Ann. New York
. Acad. Sci. 280 1976 279280.
w x 24 T.W. Deacon, The neural circuitry undelying primate calls and
. human language, Hum. Evol. 4 1989 367401.
w x 25 T.W. Deacon, Cortical connections of the inferior arcuate sulcus
. cortex in the macaque brain, Brain Res. 573 1992 826.
w x 26 D.C. Dennett, Darwins Dangerous Idea. Evolution and the Mean-
ings of Life, Simon and Schuster, New York, 1995, 586 pp.
w x 27 A.J. Doupe, A neural circuit specialized for vocal learning, Curr.
. Opin. Neurobiol 3 1993 104111.
w x 28 N.F. Dronkers, A new brain region for coordinating speech articula-
. tion, Nature 384 1996 159161.
w x 29 G.M. Edelman, Neural Darwinism. The Theory of Neural Group
Selection, Basic Books, New York, 1987, 371 pp.
w x 30 D. Falk, Braindance, Henry Holt, NY, 1992, 260 pp.
w x 31 J.A. Fodor, The Modularity of Mind, MIT Press, Cambridge, 1983,
143 pp.
w x 32 R.S.J. Frackowiak, Functional mapping of verbal memory and lan-
. guage, Trends Neurosci. 17 1994 109115.
w x . 33 J.M. Fuster, Frontal lobes, Curr. Opin. Neurobiol. 3 1993 160165.
w x 34 J.M. Fuster, Memory in the Cerebral Cortex, MIT Press, Cambridge,
1995, 358 pp.
w x 35 A.M. Galaburda, F. Sanides, Cytoarchitectonic organization of the
. human auditory cortex, J. Comp. Neurol. 190 1980 597610.
w x 36 A.M. Galaburda, M. Le May, T. Kemper, N. Geschwind, Leftright
. asymmetries in the brain, Science 199 1978 852856.
w x 37 N. Geschwind, The development of the brain and the evolution of
. language, Mon. Ser. Lang. Ling. 1 1964 155169.
w x 38 N. Geschwind, W. Levitsky, Human Brain: leftright asymmetries
. in temporal speech region, Science 161 1968 186187.
w x 39 T. Givon, Functionalism and Grammar, John Benjamin, Amsterdam,
1995, 486 pp.
w x 40 P.S. Goldman-Rakic, Development of cortical circuitry and cogni-
. tive function, Child Devel. 58 1987 642691.
w x 41 P.S. Goldman-Rakic, Circuitry of the prefrontal cortex and the
region of behavior by representational knowledge, in: F. Blum, V.
. Mountcastle Eds. , Handbook of Physiology, Am. Physiol. Society,
Bethesda, 1987, pp. 373417.
w x 42 P.S. Goldman-Rakic, Cellular basis of working memory, Neuron 14
. 1995 477485.
w x 43 P.S. Goldman-Rakic, Architecture of the prefrontal cortex and the
. central executive, Ann. NY Acad. Sci. 769 1995 7183.
w x 44 P.S. Goldman-Rakic, Regional and cellular fractionation of working
. . memory, Proc. Natl. Acad. Sci. USA 93 1996 1347313476.
w x 45 P.M. Greenfield, Language, tools and brain: the ontolgeny and
phylogeny of hierarchically organized sequential behavior, Behav.
. Brain Sci. 14 1992 531595.
w x 46 M. Habib, J.F. Demonet, Cognitive neuroanatomy of language: the
. contribution of functional neuroimaging, Aphasiology 10 1996
217234.
w x 47 H.E. Heffner, R.S. Heffner, Temporal lobe lesions and perception of
. species-specific vocalizations in macaques, Science 226 1984 75
76.
w x 48 A. Ide, E. Rodrguez, E. Zaidel, F. Aboitiz, Bifurcation patterns in
the human Sylvian fissure: hemispheric and sex differences, Cereb.
. Cortex 6 1996 717725.
w x 49 R. Jackendoff, Semantic Structures, MIT Press, Cambridge, 1990,
329 pp.
w x 50 L. Jancke, G. Schlaug, Y. Huang, H. Steinmetz, Asymmetry of the
. planum parietale, Neuroreport 5 1994 11611163.
w x 51 J.H. Kaas, The functional organization of somatosensory cortex in
. primates, Ann Anat. 175 1993 509518.
w x 52 P. Lieberman, On the evolution of human syntactic ability: its
pre-adaptive basesmotor control and speech, J. Hum. Evol. 14
. 1985 657665.
w x 53 W.C. Loftus, M.J. Tramo, C.E. Thomas, R.L. Green, N.A. Nord-
gren, M.S. Gazzaniga, Three-dimensional quantitative analysis of
hemispheric asymmetry in the human superior temporal region,
. Cereb. Cortex 3 1993 348355.
w x 54 E. Murray, M. Mishkin, Amygdalectomy impairs crossmodal associ-
. ation in monkeys, Science 228 1985 604606.
w x 55 F.K.D. Nahm, D. Tranel, H. Damasio, A.R. Damasio, Cross-modal
. associations and the human amygdala, Neuropsychologia 31 1993
727744.
w x 56 D.N. Pandya, Anatomy of the auditory cortex, Rev. Neurol. 151
. 1995 486494.
w x 57 D.N. Pandya, E.H. Yeterian, Architecture and connections of corti-
. cal association areas, in: A. Peters, E.G. Jones Eds. , Cerebral
Cortex, vol. 4, Plenum, New York, 1985, pp. 361.
w x 58 D.N. Pandya, E.H. Yeterian, Architecture and connectivity of cere-
bral cortex: implications for brain evolution and function, in: A.B.
. Scheibel, A.F. Wechsler Eds. , Neurobiology of Higher Cognitive
Function, Guilford Press, New York, 1990, pp. 5384.
w x 59 E. Paulesu, C.D. Frith, R.S.J. Frackowiak, The neural correlates of
. the verbal component of working memory, Nature 362 1993
342345.
w x 60 B.R. Payne, Evidence for visual area homologs in cat and macaque
. monkey, Cereb. Cortex 3 1993 125.
w x 61 M. Petrides, Lateral frontal cortical contribution to memory, Sem.
. Neurosci. 8 1996 5763.
w x 62 M. Petrides, D.N. Pandya, Projections to the frontal cortex from the
posterior parietal region in the rhesus monkey, J. Comp. Neurol. 228
. 1984 105116.
w x 63 M. Petrides, D.N. Pandya, Association fiber pathways to the frontal
cortex from the superior temporal region in the rhesus monkey, J.
. Comp. Neurol. 273 1988 5266.
w x 64 M. Petrides, B. Alivisatos, E. Meyer, A.C. Evans, Dissociation of
human mid-dorsolateral from posterior dorsolateral frontal cortex in
. . memory processing, Proc. Nat. Acad. Sci. USA 90 1993 873877.
w x 65 M. Petrides, B. Alivisatos, E. Meyer, A.C. Evans, Functional activa-
tion of the human frontal cortex during the performance of verbal
. . working memory tasks, Proc. Nat. Acad. Sci. USA 90 1993
878882.
w x 66 S. Pinker, P. Bloom, Natural language and natural selection, Behav.
. Brain Sci. 13 1990 707784.
w x 67 D. Premack, The codes of man and beasts, Behav. Brain Sci. 6
. 1983 125167.
w x 68 T. Preuss, The argument from humans to animals in cognitive
. neuroscience, in: M.S. Gazzaniga Ed. , The Cognitive Neuro-
sciences, MIT Press, Boston, 1995, pp. 12271241.
w x 69 T. Preuss, P.S. Goldman-Rakic, Myelo- and cytoarchitecture of the
granular frontal cortex and surrounding regions in the strepsirhine
primate Galago and the anthropoid primate Macaca, J. Comp.
. Neurol. 310 1991 429474.
w x 70 T. Preuss, P.S. Goldman-Rakic, Architectonics of the parietal and
temporal association cortex in the strepsirhine primate Galago com-
pared to the anthropoid primate Macaca, J. Comp. Neurol. 310
. 1991 475506.
w x 71 T. Preuss, P.S. Goldman-Rakic, Ipsilateral cortical connections of
granular frontal cortex in the strepsirhine primate Galago, with
comparative comments on anthropoid primates, J. Comp. Neurol.
. 310 1991 507549.
w x 72 R.A. Raff, The Shape of Life, Chicago Univ. Press, Chicago, 1996,
520 pp.
w x 73 J.P. Rauschecker, B. Tian, M. Hauser, Processing of complex sounds
. in the macaque nonprimary auditory cortex, Science 268 1995
111114.
w x 74 R.J. Richards, Darwin and the Emergence of Evolutionary Theories
of Mind and Behavior, Univ. of Chicago Press, Chicago, 1987, 700
pp.
w x 75 E. Salmon, M. Van der Linden, F. Collette, G. Delfiore, P. Maquet,
C. Degueldre, A. Luxen, G. Franck, Regional brain activity during
. working memory tasks, Brain 119 1996 16171625.
w x 76 M.S. Seidenberg, Language acquisition and use: learning and apply-
. ing probabilistic constraints, Science 275 1997 15991603.
w x 77 B. Seltzer, D.N. Pandya, Parietal, temporal, and occipital projections
to cortex of the superior temporal sulcus in the rhesus monkey: a
. retrograde tracer study, J. Comp. Neurol. 343 1994 445463.
w x 78 R.M. Seyfarth, L. Cheney, Meaning and mind in monkeys, Sci. Am.
. 267 1992 7884.
w x 79 M.J. Tovee, E.M. Cohen-Tovee, Working memory: trouble in mind,
. Curr. Biol. 6 1996 1315.
w x 80 A. Trortais, Impairment of memory for sequences in conduction
. aphasia, Neuropsychologia 12 1974 355366.
w x 81 I. Wickelgren, Getting a grasp on working memory, Science 275
. 1997 15801582.
w x 82 W.K. Wilkins, J. Wakefield, Brain evolution and neurolinguistic
. preconditions, Behav. Brain Sci. 18 1995 161181.
w x 83 F.A.W. Wilson, S.P. O Scalaide, P.S. Goldman-Rakic, Dissociation
of object and spatial processing domains in primate prefrontal
. cortex, Science 260 1993 19551958.
w x 84 S.F. Witelson, D.L. Kigar, Sylvian fissure morphology and asymme-
try in men and women: bilateral differences in relation to handed-
. ness in men, J. Comp. Neurol. 323 1992 326340.
w x 85 M.P. Young, The organization of neural systems in the primate
. cerebral cortex, Proc. Roy. Soc. London B252 1993 1318.
w x . 86 E. Zaidel, 1985 Introduction to biological and psychological stud-
. ies on hemispheric specialisation, in: D.F. Benson, E. Zaidel Eds. ,
The Dual Brain: Hemispheric Specialization in Humans, Guilford
Press, New York, pp. 4763.
w x 87 S. Zeki, S. Shipp, The functional logic of cortical connections,
. Nature 335 1988 311317.

The Evolutionary Origin of The Language Areas in The Human Brain. A Neuroanatomical Perspective

Enviado por

Dados do documento

Descrição original:

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

The Evolutionary Origin of The Language Areas in The Human Brain. A Neuroanatomical Perspective

Enviado por

Direitos autorais:

Formatos disponíveis

.

Brain Research Reviews 25 1997 381396

opercularis of the inferior frontal gyrus Brodmanns areas

in which agrammatism inability to handle and to generate

Sylvian fissure the planum temporale, that corresponds to

region in the left hemisphere other superior temporal

the inferior frontal gyrus however, connectional studies in

regions involved in language processing thalamus, basal

differentiation of the ventral premotor region ventral area

involved in working memory for object naming see Sec-

investigated. The superior temporal lobe especially on the

from area Tpt end in dorsal area 8 dorsal prearcuate

also area 6 in the dorsal frontal lobe the Tpt moiety

the inferior arcuate sulcus corresponding to Brocas re-

tent with the higher resolution based on anterograde trac-

while parietal areas 7ip and 7b project to the anterior area

connecting auditory regions with parietal areas including

not well developed. Considering that the vast majority of

macaque, the equivalent to Brocas area understood in a

area TE, while the homologue of Wernickes area area

the posterior insula which is interposed between temporal

strengthening of this temporo-parietal link see Section

includes inferoparietal and ventral premotor regions see

notable increase in relation to other mammals this is

extensive growth in the anthropoid lineage no equivalents

an immediate clue to solve the problem that is required as

before the response comes to an end for review see Refs.

between the inferior temporal lobe where object features

say, relatively well-developed cross-modal corticocorti-

preadaptation or in more modern terms, exaptation or

cross-modal associations visualsomatosensory, visual

involving both a phonological store and a subvocal re-

ment for letters involving only the subvocal rehearsal

lesions in area 40 and the arcuate fascicules presumably

From the studies in the monkey reviewed before Sec-

while the supramarginal gyrus area 40, and perhaps also

participating in higher levels of language processing see

On the other hand, granular frontal areas area 9 and

memory processes i.e., saying aloud the numbers 1 to 10

oparietal areas 3940 and frontal areas 4447 which

and the connectional information discussed above Sec-

related to working memory tasks this fits the notion of an

memory systems exists in the prefrontal cortex although

complex vocalizations was also being developed see Sec-

some basic components of syntactic rules including

involved in other cognitive tasks especially granular frontal

Brocas and Wernickes areas and parietal regions plus

mined but on the other hand syntax especially high-level

a genetic representation of such rules we are not born

origin of language this explains the linguistic abilities of

phonological rehearsal loop between the inferior parietal

jects directly to prefrontal cortex Brocas area in an

the history of primates something similar occurs with

previously established asymmetries for sequential left

Você também pode gostar