SYSTEMIC PROFESSIONAL DEVELOPMENT 2022
TEXT 2 - INTAKE
This is an extract from the book Key Questions in Second Language Acquisition by Bill
VanPatten, Megan Smith, and Alessandro G. Benati (2020, p. 46-48).
Learners are exposed to vast amounts of input over time. The child L1 acquirer at
the age of 1.5 years might hear hundreds of sentences during a typical day. The L2
learner might hear, see, or read hundreds of sentences in a given week. Yet acquisition is
slow. The child L1 learner does not get a fully developed adult-like system for years. For
the L2 learner, given limited contact with the L2, even more time (in terms of years) is
involved. Why is acquisition so slow if input is so abundant? In a seminal 1967 essay, S.
Pit Corder coined the construct intake. Intake is that subset of the input that a learner can
actually make use of at a given time. Scholars vary slightly on the exact de nition of
intake, but one de nition we think works best is this: Intake is that subset of the input for
which a learner can connect form and meaning during real-time comprehension. For
example, in the earliest stages of learning English, a learner might hear “The professor
gives way too much homework.” Using context and other non-linguistic cues, the learner
attaches meaning to “professor,” “give,” and “homework.” The rest is mere noise. It can’t
form part of intake. Later in acquisition, the learner might hear the same sentence and
this time process not just “professor,” “give,” and “homework” but also “too much.” Later,
the learner might hear the same sentence but this time process not only those same
words but also “the.” In this simple example, we see that something internal to the learner
constrains what part of the input the learner can actually process while trying to
comprehend a speech stream. While the outside world largely (but not exclusively)
controls the kind of input learners receive, learners are the ones who control the intake–
what actually gets processed at any given point in time and becomes available for
acquisition. And to be sure, the internal mechanisms and constraints on acquisition
discussed in Chapter 2 may further narrow what actually becomes part of the developing
linguistic system.
While the outside world largely controls the kind of input learners receive, learners are
the ones who control the intake–what actually gets processed.
Interestingly, while intake is a concept accepted by most if not all scholars of SLA,
little research has been conducted to determine what factors and what mechanisms are
involved in the “conversion of input to intake.” Some scholars have considered the role of
working memory–just how much information people can process at any given moment–
and how that a ects what parts of the input learners “get.” Others have looked at
linguistic constraints such that learners can’t process B until they can already process A
(e.g., in our example, learners tend not to process determiners like the and a/ an until they
can readily process nouns like professor and dog). And to be sure, there are a few other
ideas about constraints on the processing of input and we have listed some suggested
readings in the conclusion of this chapter. The point is that scholars know that learners
don’t process all of the input they are exposed to at a given point; they just have not
converged on clear explanations of why.
1
ff
fi
fi
We have mentioned the word process several times and also mentioned form–
meaning connections. Before continuing, it might be a good idea to touch on the
complexity of these notions. Above we used simple vocabulary items like linking
professor to its intended meaning and later the professor to its intended meaning. One
reason the creation of intake is so complex is that there is a lot of hidden “stu ” going on
in linking form to meaning. The simple word the, for example, has a number of features
associated with it that must get “tagged” and incorporated into the word for it to be fully
represented in the learner’s mind/ brain. For example, it must be tagged with the feature
“D” which stands for “determiner.” This feature restricts where the word can be used in a
sentence; that is, it can only occur where determiners are allowed. Without this feature
attached to it, the word could show up anywhere in a sentence! But we know it can’t. So
the learner’s mechanisms have to tag it with this syntactic feature. The learner’s internal
mechanisms must also tag it with the feature “speci c.” This abstract feature is what
distinguishes the from a/ an as well as the lack of any determiner in sentences such as
Dogs make good pets (as opposed to The dogs make good pets). Another abstract
feature that has to get tagged with the is “known to the speaker and hearer.” This is the
feature that allows the person to understand the di erence between The dogs sure are
good with the kids and Dogs sure are good with kids. In the rst sentence, the entities
(particular dogs and kids) are known to the speaker and the hearer. We can’t say that for
the second sentence. So a simple morpheme such as the is not so simple as we might
rst think when we begin to peel back the linguistic layers surrounding it. It could be, and
probably is, that intake is partly constrained by all the features and behind-the-scenes
“stu ” that have to be mapped onto words and parts of words. Learners just can’t tag
everything at once. And imagine the learner’s mechanisms trying to process an entire
sentence such as The story Anne liked is very good during the earliest stages of language
acquisition.
So far, we have touched on the idea that learners are in control of intake, that
mechanisms internal to their minds constrain what parts of the speech stream get
processed at any given point in time. Some scholars have focused on factors external to
the learner that a ect the “conversion of input to intake.” Scholars like Brian
MacWhinney, Nick Ellis, and others focus on factors such as frequency. More frequent
things in the input tend to make it into intake before less frequent or infrequent things. In
ordered development, we see that-ing is much more frequent in input than third-person-s.
This might be one (and only one) reason why -ing is acquired before third-person-s.
Another factor is reliability. Does a particular thing always serve the same function and
have the same meaning in the input? If so, it is highly reliable and is more likely to make it
into intake compared to something that is less reliable (e.g., has multiple functions or
meanings). In the end, because of the complexity of language acquisition, it is likely that
factors both internal and external to learners work in tandem in the creation of intake from
input.
Consider this …
Input that learners of languages like Spanish and French receive contains abundant
examples of gender of nouns as well as gender agreement with adjectives. For example
la casa/ la maison ‘the house,’ el libro/ le livre ‘the book’ and thus la casa blanca/ la
maison blanche ‘the white house’ and el libro blanco/ le livre blanc ‘the white book.’ Yet
we know that learners rst “get” masculine and singular and only later “get” feminine.
What do you think is happening during the processing of input by learners of these
languages? Do you think somehow they are not processing gender and gender
agreement? Do they skip it in the input?
2
fi
ff
ff
fi
ff
fi
fi
ff