Page 294-317, 24 pages Page 1 of11
Chapter 11: Language
WHAT IS LANGUAGE?
• Language: A system of communication using sounds or symbols that enables us to express our
feelings, thoughts, ideas, and experiences in either spoken or written form. This allows the remarkable
feat of communication!
• This alone is too broad for human language, can apply to animal calls or behaviours like bee dances.
The Creativity of Human Language
• Language provides a way of arranging a sequence of signals (sounds, letters, words) to transmit things
ranging from simple and commonplace, to messages that perhaps have never been previously
composed in the history of the world can create new and unique sentences
• This is enabled by its hierarchical structure, where language consists of a series of small components
that can be combined to form larger units, e.g. words to phrases, to sentences, to stories.
• Language is also governed by rules, such that these components can be arranged in certain ways but
• This goes beyond the fixed calls and signs of animals, that transmit single messages such as “feed me”
The Universality of Language
• The need to communicate is universal. Everyone with normal capacities develops a language and
learns to follow its complex rules, even though they may not be aware of these rules consciously. Deaf
children in an environment without sign language will invent a coherent sign language themselves!
• Language is universal across cultures; there is not a single culture without language, totalling 5000
• Language development is similar across cultures, with babbling at 7 months, some meaningful words
around age 1, and multiword utterances at age 2. There is a critical period, with loss of ability to acquire
fluency in a new language after age 7 or so; vocal cords change to cater to the language we are
• “Unique but the same” with different words, sounds, and rules but all have similar functions with ability
to ask questions, to refer to the past and present, to make things negative, etc.
• Broca and Wernicke involved in early research of brain areas involved in understanding and producing
language, but today we know language processing is more distributed.
• B.F. Skinner proposed that language was learned through reinforcement, as part of behaviourism, with
children’s correct language use being rewarded.
• Noam Chomsky thought that human language is innate and coded in the genome, that despite wide
variations that exist across languages the underlying basis is similar. Evidence against behaviourism that
as children learn language, they produce sentences that they have been taught or reinforced by parents
(“I hate you, mommy!”)
• Psycholinguistics: Field of psychological study of language. Concerned with:
o Comprehension: How do people understand language, e.g. process language sounds,
understand words in writing, have conversations.
o Speech Production: How do people produce language, both physically and mentally?
o Representation: How is language represented in the mind, to make connections between
different words and sentences for cohesion? Page 294-317, 24 pages Page 2 of11
o Acquisition: How do people learn language, either as children or later in life?
PERCEIVING WORDS, PHONEMES, AND LETTERS
• Infants rapidly acquire words in their second year, and adults understand more than 50,000 words!
• Lexicon: A person’s knowledge of what words mean, how they sound, and how they are used. The
more languages one speaks, the more lexicons one has.
Components of Words
• Phoneme is the shortest sound segment of speech that, if changed, changes the meaning of a word.
E.g. bit has phonemes /b/, /i/, and /t/. If /b/ is changed to /p/, the word is now pit.
• Phonemes ARE NOT letters, which can each have a number of different sounds (“e” in “we” and “wet”)
or even be silent (“e” in “some”).
• The number of phonemes varies in different languages: 47 in English, 60 in some African dialects
• Morpheme is the smallest unit of language with a definable meaning or grammatical function. For
example, “truck” has a few phonemes, but only one morpheme.
• In contrast “bedroom” has two morphemes, “bed” and “room”. Endings such as “s” in “trucks” are also
morphemes since it indicates more than one.
Perceiving Spoken Phonemes, Words, & Written Letters
• An important characteristic of language is that its various components are affected by the context within
which they are heard or seen. For this to occur, we must first perceive phonemes, words, and meanings.
Speech: Perceiving Phonemes
• Warren (1970): Phonemic Restoration Effect, where a phoneme
that is part of a sentence can be heard even if its sound is covered up
by extraneous noise. This is because the missing phoneme is filled in
by the context produced by the sentence and the rest of the world, in
• Subjects that hear “The state governors met with their legislatures
convening in the capital city”, but with first /s/ in “legislatures” covered
by sound of a cough. The subjects did not even notice the /s/ was
missing, and could not remember where the cough occurred.
• This effect can thus be influenced by changing the context. “There
was time to *ave” could be shave, save, wave, etc. but participants
report hearing “wave” when the remainder of the sentence had to do with saying goodbye. Thus our
knowledge and expectations affect speech perception.
• More restoration occurs for a real world like prOgress than a pseudoword like crOgress for which we do
not have contextual knowledge
Speech: Perceiving Words
• People talk with different accents, different enunciations, and at different speeds. This makes it difficult
to interpret or perceive words especially when taken out of context.
• Pollack & Pickett (1964): Recorded conversations of subjects when in waiting room. The subjects are
then presented with single words taken out of their own conversations; they could only identify half the
words! This reinforces the idea that word perception requires context. Page 294-317, 24 pages Page 3 of 11
• Speech Segmentation: The process of perceiving individual words in the continuous flow of the
speech signal. In conversation, words are not separated from each other by spaces or pauses, although
it may sound like they are, when we look at the physical sound energy from a sentence.
• Speech segmentation therefore does not occur through perceiving actual pauses in speech. It relies on
knowledge of word meanings to separate words; we cannot distinguish words in an unfamiliar foreign
language. For example, in “I scream, you scream, we all scream for ice cream” the sound stimuli for “I
scream” and “ice cream” are actually identical and require understanding context for the correct
• Segmentation is also aided by knowledge about rules: how likely certain sounds are to follow one
another within a word, vs. being separated into two words. E.g. prettybaby with pre and ty known to be
more likely to be in one word, and ty and ba in two different words to correctly insert the space to form
Reading: Perceiving Letters
• Word Superiority Effect: Context is also important for perceiving
written letters. In this effect, letters are easier to recognize when they
are contained in a word than when they appear alone or are in a
• Reicher (1969): A stimulus that is either 1) a word (FORK), 2) a single
letter (K), or a nonword (RFOK) is flashed. It is followed by a random
pattern where the stimulus was, and two letters – one that appeared in
the original stimulus (K) and one that didn’t (M).
• The subject must pick the letter that was presented in the stimulus.
Subjects picked the correct answer more quickly and accurately when
the letter had been part of a word, than if alone, than if in a nonword.
• So letters are not processed one by one, rather each letter is affected by its surroundings (top-
The Word Frequency Effect
• Word Frequency: The relative usage of a word in a particular language, e.g. home 547/million words
while hike only 4/million. Word Frequency Effect: The fact that we respond more rapidly to high-
frequency words than low-frequency words.
• Lexical Decision Task: Involves reading a list that consists of words and nonwords, task is to indicate
as quickly as possible whether each entry is a word or not. This decision about whether each word was
present in your lexicon is a lexical decision task.
• Savin (1963): People read high-frequency words faster and carry out the lexical decision task more
rapidly than for low-frequency words.
• Rayner (2003): This is also demonstrated through tracking people’s eye movements during reading.
Readers fixate on low-frequency words 40ms longer than high, perhaps because they need more time to
access the meaning of low-frequency words. Past experience with words influence our ability to
access word meaning.
• Lexical Ambiguity: Words that have more than one meaning, e.g. bug as insects, hidden listening
devices, or to annoy. When this occurs, we use the sentence context to determine which definition
applies, usually without conscious awareness.
• Swinney (1979): People briefly access multiple meanings of ambiguous words before context takes
over. Subjects are presented with a ecording of sentences like “spiders, roaches, and other bugs” where
bug = insects. Page 294-317, 24 pages Page 4 of 11
• Lexical Priming: In priming, presentation of first stimulus activates a representation of it, so a person
can respond more rapidly to the stimulus when it is presented again. Here, lexical priming involves the
meaning of words where priming with one word enhances response to another word with a similar
meaning – e.g. ant and bug
• In this experiment, as subjects hear the word bug, they are visually presented with either a word or
nonword on a screen; the lexical decision task is to decide whether it is a word or nonword.
• The word could either be related to the insect meaning (ant), the hidden listening device (spy), or not
related at all (sky). Results show that participants responded with similar speed to both ant and spy,
significantly faster than to sky. This means during presentation listeners accessed both “listening device”
and “insect” meanings of bug
• The effect is no longer seen when the test is done 200ms after bug is heard – at this point, the “insect”
meaning has been selected and the “listening device” meaning dropped/inactivated.
• Context thus exerts its influence to determine the correct meaning after a slight delay, during
which other meanings of a word are briefly accessed
• Semantics: The meaning of words and sentences. Syntax: The rules for combining words into
sentences, including grammar and word order, e.g. “The cat chased the bird” vs. “Cat bird chased the”
Parsing & a Trip down the Garden Path
• Parsing: The grouping of words into phrases, for determining the meaning of a sentence
• Temporal Ambiguity: Sentences that have more than one meaning, specifically in that the initial words
of a sentence can lead to more than one consequence, e.g. “Amanda believed the senator…” to “during
his speech” vs. “was lying to the committee” – two opposite meanings of senator telling the truth or lying
• Individuals typically decide on meanings as the sentence unfolds, not only when the sentence is
completed; this means one may have to revise one’s understanding from an earlier decision
• Garden Path Sentence: It leads the reader down a path that seems right, but turns out to be wrong,
due to differences in parsing. e.g. “Cast iron sinks quickly rust” where one first parses “cast iron sink” as
an iron sink in a kitchen, but then “cast iron sinks quickly” as metal sinking in water, then back to a
kitchen sink again with “rust”
• E.g. “The man who whistles tunes pianos”, or “The man who hunts ducks out on weekends”
The Syntax-First Approach to Parsing
• Syntax-First Approach to Parsing: Parsing is determined by the grammatical structure of the
sentence, grouping based on structural principles such as late closure, where when one encounters a
new word, the person’s parsing mechanism assumes this word is part of the current phrase such that
each new word is added to the current phrase as long as possible.
o |Cast iron| indicates a type of iron
o |Cast iron sinks| as either kitchen sinks or metal sinking to the bottom (lexical ambiguity
o |Cast iron sinks quickly| with “quickly” added to current phrase in late closure
o Finally, |cast iron sinks||quickly rust| where “rust” makes it necessary to adjust the parsing so
“quickly” becomes part of a new phrase – late closure had led us down the garden path
o Syntax-based late closure controls original interpretation (leads reader astray), but
correction is made after the real meaning becomes clear with semantics rearranging the
parsing. Page 294-317, 24 pages Page 5 of11
The Interactionist Approach to Parsing
• Interactionist Approach to Parsing: All information, both syntactic and
semantic, is taken into account simultaneously as we read or listen to a
sentence, so any corrections take place as the sentence is unfolding
Sentence Understanding Influenced by Meanings of Words
• “The spy saw the man with the binoculars” has two meanings, either
“The spy with the binoculars is looking at the man”, or “The spy is looking
at a man, who has a pair of binoculars”
• But if we change just one word, only one meaning becomes reasonable:
“The bird saw the man with the binoculars”. The interpretation where the
bird uses the binoculars is not even considered; although this sentence
has the same syntactical structure as the spy sentence, our knowledge of the properties of spies and
birds influence the way we interpret the relationships between the words in the sentence
• Thus semantics can be important in determining parsing right at the beginning of the sentence
Sentence Understanding Influenced by Environment Setting
• Our interpretation is also influenced by the meaning of the scene we may be observing.
• Tanenhaus (1995): Subjects are presented with objects on a table, either one-apple condition or two-
apple condition with a second apple on a napkin. This display is visible to the subject as they listen to the
instruction “Put the apple on the towel in the box”. Subject eye movements as they listen to the
instructions are tracked.
o In one-apple condition: At
first sounds like the apple
should be moved to the other
towel, but then corrected that
it should be placed in the box.
Eye tracking shows that
person first looks at the apple
(1), then at the other towel (2)
[initial interpretation], then at
the apple (3) and the box (4)
with the correct interpretation
o In two-apple condition:
First at apple on napkin (1), then at the apple on the towel (2), then at the box (3)
o ^ On the towel is interpreted as not indicating whether an apple should be placed on the other
towel, but as the apple which should be moved
• Syntax-first approach would have predicted initial understanding being that the apple is to be placed on
the towel – fails for the two-apple condition. The fact that a different result occurs in the two-apple
condition means the listener is taking both the syntactic information and the scene information
• Thus information in addition to the structure of the sentence help determine meaning; we rarely
encounter sentences in isolation, but rather in context of specific environments, conversations, stories,
etc. We have a great deal of information that we can apply to a specific sentence
• Discourse/text processing is the study of how we understand text and stories
UNDERSTANDING TEXT & STORIES
• Sentences in one part of the story related to sentences in other parts, such that these relationships
create a coherent, understandable story. This involves inferences to determine meaning, a creative
process involving pri