Language and Brain

What are the neurobiological precursors of language? How is language development related to the development of
brain functions? How do biological processes interact with social factors?

Are language functions localized?

The brain is not a homogeneous mass; parts of it are specialized for specific tasks. How do
we know this? In the past most of our knowledge about how brain and behavior are related
came from lesion studies combined with an autopsy: neuropsychologists would discover which
part of the brain had been damaged, and relate that information to behavior. Now we have
brain imaging techniques available, particularly fMRI ,which can also be used with non-brain-
damaged speakers. These techniques indicate which parts of the brain are active when we do
tasks such as reading or speaking. Most people know that the brain is divided into two
hemispheres (see Kolb & Whishaw, 2009). The two hemispheres of the brain are partly
specialized for different tasks: broadly speaking, in most right handed people the left
hemisphere is particularly concerned with analytic, time-based processing, while the right
hemisphere is particularly concerned with holistic, spatially based processing.
Localization of language
How do we know which bits of the brain do what? In the 1950s, Penfield and Roberts (1959) studied the effects of
electrical stimulation directly on the brains of patients undergoing surgical treatment for epilepsy. More recently, a
number of techniques for brain imaging have become available, including PET and CAT scans (see Chapter 1). These
techniques all show that there are specific parts of the brain responsible for specific language processes.
Most of the evidence on the localization of language functions comes from studies of the effects of brain damage.
An impairment in language production or comprehension as a result of brain damage is called aphasia. The French
neurologist Paul Broca carried out some of the earliest and most famous work on the effects of brain damage
on behavior in the 1860s. Broca observed several patients where damage to the cortex of the left frontal lobes resulted
in an impairment in the ability to speak, despite the vocal apparatus remaining intact and the ability to understand
language apparently remaining unaffected. This pattern of behavior, or syndrome, has become known as Broca’s
aphasia, and the part of the brain that Broca identified as responsible for speech production has become known as
Broca’s area (see Figure 3.6).
A few years later, in 1874, the German neurologist Carl Wernicke identified another area of the brain involved
in language, this time further back in the left hemisphere, in the part of the temporal lobe known as the temporal
gyrus. Damage to Wernicke’s area (Figure 3.7) results in Wernicke’s aphasia, characterized by fluent language that
makes little sense, and a great impairment in the ability to comprehend language, although hearing is unaffected.

The Wernicke–Geschwind model

Wernicke also advanced one of the first models of how language is organized in the brain. He argued that the “sound
images” of object names are stored in Wernicke’s area of the left upper temporal cortex of the brain. When we speak,
this information is sent along a pathway of fibers known as the arcuate fasciculus to Broca’s
FIGURE 3.6 Location of Broca’s area.

FIGURE 3.7 Location of Wernicke’s area.

area, in the left lower frontal cortex, where these sound images are translated into movements for controlling speech.
Although modern models are more detailed, they essentially still follow Wernicke’s scheme. The Wernicke–
Geschwind model (Figure 3.8; sometimes called the Wernicke–Lichtheim–Geschwind model) is an elaboration of
Wernicke’s scheme. Geschwind (1972) described how language generation flows from areas at the back to the front
of the left hemisphere. When we hear a word, information is transmitted from the part of the cortex responsible for
processing auditory information to Wernicke’s area. If we then speak that word, information flows to Broca’s area
where articulatory information is activated, and is then passed on to the motor area responsible for speech. If the word
is to be spelled out, the auditory pattern is transmitted to a structure known as the angular gyrus. If we read a word,
the visual area of the cortex activates the angular gyrus and then Wernicke’s area. Wernicke’s area plays a central
role in language comprehension. Damage to the arcuate fasciculus results in difficulties repeating language, while
comprehension and production remain otherwise unimpaired. This pattern is an example of a disconnection
syndrome. Disconnection occurs when the connection between two areas of the brain is damaged without damage to
the areas themselves. The angular gyrus plays a central role in mediating between visual and auditory language.
 Some researchers have suggested that the right hemisphere plays an important role in an acquired disorder
known as deep dyslexia (also known as a learning disorder, difficulty in recognition speech sounds leading to
difficulty in reading and spelling out words accurately) (see Chapter 7), that it carries out important aspects of visual
word recognition, and that it is involved with aspects of speech production, particularly prosody (regarding the
loudness, rhythm, pitch, and intonation of speech); see Lindell (2006) for a review.

Subcortical regions of the brain might play a role in rule based aspects of the language. For example, Ullman
et al. (1997) found that although people with Parkinson’s disease (which affects subcortical regions of the brain)
could successfully inflect irregular verbs (presumably because these are stored as specific instances rather than
generated by a rule), they had difficulty with regular verbs, suggesting that subcortical regions play some role in rule-
based aspects of language.

The right cerebellum becomes significantly activated when we process the meaning of words (Marien,
Enggelborghs, Fabbro, & De Deyn, 2001; Noppeny & Price, 2002; Paquier & Marien, 2005; Petersen, van Mier, Fiez,
& Raichle, 1998). Second, even within the left cortex it is clear that brain regions outside the traditional Wernicke–
Broca areas play an important role in language. In particular, the whole of the superior temporal gyrus (of which
Wernicke’s region is just part) is important. Third, brain damage does not have such a clear-cut effect as the model
predicts. Complete destruction of areas central to the model rarely results in permanent aphasias of the expected types.
Furthermore, we rarely find the expected clear-cut distinction between expressive (production) and receptive
(comprehension) disorders. For example, people with damage to Broca’s region often have difficulty understanding
sentences. Different types of aphasia have variable clusters of symptoms that tend to go together, and that are not as
clearly related to regions such as Broca’s or Wernicke’s as the model predicts. Fourth, virtually all people with aphasia
have some anomia (difficulty in finding the names of things) regardless of the site of damage. Finally, electrical
stimulation of different regions of the brain often has the same effect, and selective stimulation of Broca’s and
Wernicke’s areas does not produce the simple, different effects that we might expect.

More recent models of how language is related to the brain

Ullman (2004) proposed a model, called the D/P (declarative/procedural) model, of how language relates to the brain.
He argued that language depends on two brain systems. The mental dictionary, or lexicon, depends on a
declarative memory system based mainly in the left temporal lobe. The mental grammar, which depends
primarily on procedural memory, is based on a distinct neural system involving the frontal lobes, basal ganglia,
cerebellum, and regions of the left parietal lobe. Essentially this distinction is one between linguistic rules, or
syntax, and words. The distinction will recur throughout this book, so it is important to remember that there is some
anatomical
Colored PET scans of the
areas of the brain active
while understanding
language. (The fronts of
these human brains are to
the left.) Active parts of
the cerebral left
hemispheres are
red/orange. Various
language areas of the extra
sylvian temporal region are
active in the scan on the
left, which shows activity
associated with working
out the meaning of words.
The scan on the right
shows activity associated
with understanding
sentences.

justification for this distinction. Another important idea here is that language processing makes some use of cognitive
processes and brain structures that are not just dedicated to language.
Recent work has used imaging to explore the exact role of Broca’s area in language, and one result is that its
precise role has become much more controversial. The fact that damage to Broca’s area leads to aphasia shows that
it plays an important role, but is it dedicated to language specifically, or does it just involve more general processes
that underpin language? Are other regions of the brain involved in processing syntax? The answer to the latter question
is almost certainly yes, and to the former, maybe. Imaging suggests that Broca’s area may play a role in general
phonological working memory rather than syntactic manipulation as such (Rogalsky & Hickok, 2011; but see also
Fedorenko & Kanwisher, 2011). There is even debate as to the exact language-related processes that Broca’s area
computes, including phonological short-term memory (Rogalsky & Hickok, 2011), building a hierarchical structure
(Friederici, 2002; Friederici, Bahlmann, Heim, Schubotz, & Anwander, 2006), linearizing a hierarchical structure
(Bornkessel-Schlesewsky, Schlesewsky, & von Cramon, 2009), and unifying concepts into a planned sentence
(Hagoort, 2008). Quite a list!
FIGURE 3.9 (A) Hickok and Poeppel’s proposed framework for the functional anatomy of language. (B) General
locations of the model components shown on a lateral view of the brain. From Hickok and Poeppel (2004).
Some portions of the brain are more important for language functions than others, but it is difficult to localize
specific processes in specific brain structures or areas. It is likely that multiple routes in the brain are involved in
language production and comprehension. Modern brain-imaging techniques show that much larger regions of the
brain may be involved in language processing than were once thought. For example, the temporal gyrus seems to play
an important role in language comprehension (Dronkers, Wilkins, van Valin, Redfern, & Jaeger, 2004). A wide-
ranging account of the relation between language and the brain is provided by Hickok and Poeppel (2004), who,
drawing on data from brain imaging and lesion studies, focus on auditory comprehension. They argue that early stages
of speech perception involve the superior temporal gyrus bilaterally (on both sides, although more on the left). The
cortical processing system then diverges into dorsal (towards the back and top of the brain) and ventral (towards the
front and bottom of the brain) streams (see Figure 3.9). The ventral stream is mainly concerned with turning sound
into meaning. The dorsal stream is concerned with mapping sound onto a representation involving articulation, and
relates speech perception to speech production. Most of what we traditionally think of as “speech perception” takes
place in the ventral stream. The output of the dorsal stream is an integration of auditory and motor information, and
the stream is important when we focus on the sounds of the words involved (e.g., in learning to make speech sounds,
or in analyzing the sounds of words, or repeating back nonwords).
 In summary, although we can point to specific regions of the brain—particularly in the left frontal and temporal
lobes—that play particularly important roles in language, lesion and imaging studies show that the neural systems
underlying language are variable, flexible, and distributed over many brain regions (Corina et al., 2003).
In a recent synthesis, Friederici (2012) describes how the cortical regions of the brain involved in language are
connected by ventral and dorsal pathways. The ventral pathway is involved in auditory-to-meaning mapping, and the
dorsal pathway is involved in auditory-to-motor mapping. The dorsal pathway might also be involved in syntactic
processing, particularly with syntactically complex sentences. She argues that these two functions are so dissimilar
that we distinguish two dorsal streams on the basis of function and structure. The ventral pathway supports sound-to-
meaning mapping and local syntactic structure building.