34 • CHAPTER 2 Cognitive Neuroscience
The straightforward link between language production and Broca’s area and lan-
guage understanding and Wernicke’s area was for many years the accepted model of
language processing. But as we described in our introduction of models in Chapter 1
(see page 17), models are often revised in response to new data, and the Broca/Wernicke
model is no exception.
Beginning in the 1970s, researchers began providing new evidence about lan-
guage processing and the brain. One line of evidence shows how important it is to
pay close attention to how the behavior of brain-damaged patients is tested. Broca’s
idea that patients with Broca’s aphasia could understand language but had a problem
producing it has been challenged by research showing that these patients do, in fact,
have problems understanding language. Consider, for example, the following two
sentences:(1) The apple was eaten by the girl.
(2) The boy was pushed by the girl.Patients with Broca’s aphasia have no trouble understanding the fi rst sentence, but
have diffi culty with the second one. The problem they have with the second sentence
is deciding who was doing the pushing and who got pushed. Did the girl push the boy,
or did the boy push the girl? While you may think it is obvious that the girl pushed the
boy, patients with Broca’s aphasia have diffi culty processing connecting words such as
“was” and “by,” and this makes it diffi cult to determine who was pushed (notice what
happens to the sentence when these two words are omitted). In contrast, the fi rst sen-
tence cannot be interpreted in two ways. It is clear that the girl ate the apple, because
it is not possible, outside of an unlikely science fi ction scenario, for the apple to eat the
girl (Dick et al., 2001; Novick et al., 2005).
The fact that Broca’s patients do have a problem understanding language indicates
that Broca’s aphasia is not simply a problem with producing language. The results
of many behavioral and physiological experiments have caused some researchers to
distinguish not between problems of production and understanding, but between
problems of form and meaning. Form problems involve diffi culties in determining the
relation between words in a sentence (like the Broca’s aphasia patients’ problem with
sentence 2, above). Meaning problems involve wider differences in understanding like
those experienced by Wernicke’s aphasia patients, who would also have diffi culty with
sentence 1.
A method of recording rapid electrical responses of the human brain, called the
event-related potential (ERP), has provided additional evidence for distinguishing
between form and meaning in language.The event-related potential (ERP) is recorded with small disc electrodes placed on a person’s
scalp, as shown in ● Figure 2.13a. Each electrode picks up signals from groups of neurons that
fi re together. Figure 2.13b shows an event-related potential recorded as a person listens to the
phrase “The cats won’t eat.” Notice that the signals are very rapid, occurring on a time scale
of fractions of a second. This makes the ERP ideal for investigating a process such as under-
standing a conversation, in which speakers say three words per second, on the average (Levelt,
1999). The rapid response of the ERP contrasts with the slow response of brain imaging tech-
niques such as fMRI, which take seconds to develop. A disadvantage of the ERP is that it is
diffi cult to pinpoint where the response is originating in the brain. There are ways to estimate
where an ERP is originating, but it isn’t as straightforward as the fMRI, which highlights specifi c
structures that are activated. However, the ability of the ERP to provide a nearly continuous
record of what is happening in the brain from moment to moment makes it particularly well
suited for studying dynamic processes such as language (Kim & Osterhout, 2005; Osterhout
et al., in press).Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.