1. Introduced by: Supervised by : Asst. Prof. Dr.
Qassim Obaes Al-Azzawi
Hussam Kamel
Waleed K. Abdulabbas
Azhar Hanoon
2. 1. Properties of Spoken Language Production
Language comprehension is more difficult to study than language production,
because although speech is observable, the ideas that lead to production are
more elusive.
Researchers have responded to this dilemma by using convergent measures.
Some investigators have made detailed and systematic analyses of naturally
occurring errors of production, and others have given speakers under laboratory
conditions, more or less specific instructions on what to produce(Carrol,
2008:193).
3. Production process
Language production is divided into three major steps:
deciding what to express (conceptualization),
determining how to express it ( formulation)
expressing it(articulation) ( Levelt, 1989).
5. Generating : First Phase
The simplest utterance consists of a single word. Generating a word begins
with specifying its semantic and pragmatic properties. That is, a speaker
decides upon an intention or some content to express (e.g., a desired
outcome or an observation) and encodes the situational constraints on
how the content may be expressed (e.g., polite or informal speech,
monolingual or mixing languages.
This process, termed conceptualization or message planning, is
traditionally considered pre-linguistic and language neutral .
The next major stage is formulation, which in turn is divided into a word
selection stage and a sound processing stage.
6. Process of generation: Second phase
Deciding word to use involves selecting a word in one 's vocabulary based on its
semantic and pragmatic specifications. Word representation is lemma , lexical
entry, lexical representation, and it marks the presence of a word in a speaker’s
vocabulary that is capable of expressing particular semantic and pragmatic
content within a particular syntactic context.
Sound processing, involves constructing phonological form of a selected word
by retrieving its individual sounds and organizing them into stressed and
unstressed syllables (phonological encoding) and then specifying the motor
programs to realize those syllables (phonetic encoding).
The final process is articulation, that is, the execution of motor programs to
pronounce the sounds of a word.
7. Property 1: word selection precedes sound assembly
Speakers first focus on meaning and then focus on assembling their
sounds.
speaker may err in selecting a word but correctly assemble and pronounce
its component sounds.
speakers produce words first by processing their meaning-level properties,
then by processing their sound-level properties.
This implies that it is not possible to go from meaning to sounds except
through a lexical representation.
8. Property 2: The intention to produce a word activates a family of meaning
How meaning is represented in models of word production, leads to two major
theoretical positions. Decompositional views portray the primitives of semantic
representation as being entities that are smaller than the words. Put another, a
word is a composite of features e.g., the meaning of bird might include HAS
WINGS, HAS FEATHERS, SINGS SONGS, and the like.
According to non-decompositional views, the representational bases of
words and their meanings bear a one-to-one relationship, so that the word bird
is fed by an atomic meaning representation of BIRD, the word airplane is fed by
an atomic meaning representation for AIRPLANE, and so forth. These atomic
meaning representations are often called lexical concepts.
9. Property 3:Words that express similar
meanings compete for selection
when a speaker attempts to generate the most appropriate word for a
particular occasion, other words with very similar meanings in the given
context become available and compete for selection as well.
10. . Property 4: Competition for selection is constrained by
grammatical class and contextual features
Models of word production typically invoke different
processing mechanisms to impose syntactic
constraints on word selection as opposed to semantic
constraints.
, the competition is limited to words of the same
grammatical class. That is, only nouns substitute for
nouns, verbs for verbs, and so on.
11. . Property 5: The speed and accuracy of selection is affected by
meaning-level properties
The word is selected and produced are affected by a
number of factors related to semantic representations.
words with concrete, imaginable meanings such as
vampire, wind, and pine have richer semantic
representations that guide word selection more efficiently
than words with more abstract meanings such as fear,
sense, and spirit. Higher imageability or greater
concreteness facilitates word translation , generating
associations , and word recall.
12. Property 6: Attended objects do not
necessarily lead to lexical activation
word representations might be easily activated based on
any strongly associated stimulus in the environment, even
objects. For example, seeing a banana might activate the
word banana to some extent, even in the absence of any
intention to talk about it.
In contrast, viewing a semantically related object
facilitates production relative to viewing an unrelated
object (e.g., viewing a banana and generate apple)
13. . Property 7: Selecting words is subject to long-
term repetition effects that resemble learning
Selecting a word to name an object or express a meaning has long-lasting
effects on how easy it is to retrieve that word again to express a similar
meaning.
expressing the same meaning with the same word should subsequently be
more efficient, but expressing a similar meaning with a different word
should be more difficult.
14. Property 8: Word production can halt part of the way through the
process
Researchers refer to this as being in a tip-of-the-tongue or TOT state, from
the expression “to have a word on the tip of one’s tongue.” TOT states
occur most often for people’s names. Some forms of brain damage cause
people to experience similar word retrieval problems for common words.
15. Property: 9 The selection of some words and morphemes is not primarily
driven by meaning.
Generating function words and morphemes
The production of function words such as articles (a, the) and morphemes
such as past tense –ed, however, seems less driven by properties of meaning
and highly dependent on the grammatical and phonological properties of
accompanying words. For example, in English, the forms of indefinite
determiners vary depending on whether they are used with count or mass
nouns (e.g., some pasta vs. a noodle).
Other evidence also points to an important dissociation between content-
and function-word production.
16. Assembling the sounds of a word
Property 10: The sounds of a word are assembled anew
The individual sounds of words are assembled anew each time they are
spoken rather than retrieved as intact wholes.
One reason that a word’s sounds must be assembled anew each time is
due to changes in metrical structure contingent on the accompanying
words and inflections. For example, hand vs. hand it. (han –dit)
17. Property 11 : Experience strongly affects speed and accuracy of
assembling words
The sounds that slip tend to be those in the least predictable positions
within the language of the speaker. For example, word initial consonants
(e.g., /b/ in the word bicycle: /baj.sI.kl/) are less predictable than other
consonants that begin syllables in English (e.g., the /s/ and /k/ in bicycle).
children tend to learn common words earlier than uncommon words.
the impact of frequency and age-of-acquisition is greater in phonological
encoding than in word selection.
18. Property 12: Aspects of sound assembly proceed sequentially
Processing seems to start earlier in time for sounds at the beginning of
words than for sounds at the ends.
example: an initially overlapping word like tile facilitates naming of a tiger
at earlier points in time relative to a word like liar that only overlaps in
final sounds.
19. Property 13: The effect of similar sounding words is highly situation-
dependent.
Words that share many sounds with other words take less time to
generate than words that are more unusual and appear more likely to be
successfully retrieved, but
There are situations in which having similar sounding words slows
speech or increases the likelihood of errors. The most obvious case of such
interference is in tongue twisters such as The sixth sick sheik’s sixth
sheep’s sick
20. Time course of processes in word production
property 14: Semantic competitors activate their sounds
Two models are debated :
1- speakers first use activated meaning-level representations to perform word
selection and only access sound information after the completion of the selection
process.
2- To flow relatively freely among meaning, lexical, and sound representations.
• The basic idea is that the more a substituting word resembles an intended word,
the less likely a pre-articulatory editing mechanism is to detect the error and
prevent it from being uttered.
21. Property 15: The scope of message planning is greater than the scope of sound
assembly
speakers are more likely to pause, repeat words, and say um before
articulating complex utterances and clauses than less complex ones.
Although speakers seem to know a lot about the message content of an
utterance before they begin to speak, they often do not know all of the
words they will use to express the message before they begin to articulate
the utterance.
22. Syntax and Production
Syntax allows words to be combined to create unique combinations
of meaning.
A word such as cat has a particular meaning, but the expressive
power of language is enhanced by our ability to create meanings
compositionally, by putting words together – for example, our ability
to say not a cat or that’s my cat. Consider the model referred to as
Bock–Levelt (BL) as shown below
24. The Process of Syntactic
For example, for the utterance
my cat terrifies the dog next door,
the lemmas for CAT, DOG, NEXT, DOOR, and TERRIFY would be
retrieved, and CAT would be assigned the role of subject and DOG
(modified by NEXT DOOR) the role of object. But notice that a structure
such as
the dog next door, my cat terrifies him is still possible, because in this
form CAT is the grammatical subject and DOG is the object (dog is the
object in both the proposed position and in the pronominal form him).
25. Evidence for Two-Stage Models
evidence for the two-stage model is that if a speaker produces or even
simply hears an utterance with a particular structural form, he or she is
likely to mimic that structure in a subsequent utterance.
Example:
The driver showed the mechanic the overalls.
26. Is Grammatical Encoding Automatic?
This finding suggests that the process of creating syntactic
structure is resource-demanding and it might be that
grammatical encoding is automatic but semantic processing
requires planning and can therefore be resource demanding.
Are Some Constructions Difficult to Generate?
This question has received surprisingly little attention
from experimental psycholinguists.
Incremental Production
Incrementality is viewed as optimizing the use of
processing resources
27. Choice of Syntactic Construction
speakers tend to begin their sentences with that primed NP.
passives tend to occur with theme-experiencer verbs
Inclusion of Optional Functional Elements such as: The weary
traveler claimed (that) his luggage had been stolen?
Syntax and Prosody in Production
Utterances have regular rhythmic and tonal properties
Consider the sentence
If she goes, I go too.
28. Section three: Speech disorder
psycholinguistics is mainly concerned with the processes underlying the
production, perception, and comprehension of language, of which speech
is an uncontroversial component.
Psycholinguistics has an explanatory rather than a descriptive role in
speech disorders
post-stroke patient who produces an apparent speech error of replacing a
voiceless stop consonant (p, t, or k) with its voiced cognate (b, d, or g)
might be said to produce a phonemic not a phonetic error.
29. 2. Speech Delay
. Residual errors are usually thought to be the consequence of a delay in
speech motor maturation. The sounds mentioned above are said to be
residual errors because they require the most complex articulatory skill
and hence they are mastered relatively late.
/r/, /l/, and /s/, salient phonemes that are associated with speech errors.
30. 2.1. From Functional Misarticulation to
Optimality Theory
phonological processes as a key player in normal speech sound
development as well as in many cases of speech delay. A
phonological process is a psycholinguistic operation applied to the
underlying representation, the output of which is a surface form.
When a child produces open-syllable forms for words with final
consonants (e.g., [do] for dog, [kᴂ] for cat, [mi] for miss), a process
of final consonant deletion is said to intervene between the (correct)
underlying form and the surface representation.
31. Optimality theory
Optimality Theory (OT) has taken the place of phonological processes in
explaining errors and their treatments in speech delay. OT presents a set
of constraints operating between underlying forms (inputs) and surface
representations (outputs).
The advantage of OT in understanding phonological development and
disorders is best exemplified in the way the theoretical apparatus can
explain error co-occurrences and their implicational relations. This
explanation, in turn, points to a principled selection of treatment targets
having the greatest potential for generalization to untreated sounds.
32. 2.2. Consistency and Modifiability of Errors: Does the
Theoretical Viewpoint Matter?
The advantage of OT in understanding phonological development and
disorders is best exemplified in the way the theoretical apparatus can
explain error co-occurrences and their implicational relations. This
explanation, in turn, points to a principled selection of treatment targets
having the greatest potential for generalization to untreated sounds.
33. 3. Voice onset time VOT and the Voicing Distinction in Speech
Disorders
VOT is a consistently used measure in the history of speech production
research on both normal speakers and speakers with disorders.
34. Advantages of VOT
Speakers who have degenerative or acquired neurological disease are known
to produce voicing errors for stop consonants.
35. Functional Neuroimaging of Speech Production fMRI
-Functional brain imaging has proven to be a uniquely effective means to
study the role of cortical and subcortical brain structures in the production
and understanding of language.
-It has allowed examination of the regional functional specialization of
these phenomena with an accuracy and precision not previously possible
with chronometric or electrophysiological techniques.
provided a more convenient and completely non-invasive means to study
perception, language and action.
36. Acoustic Noise Effects in Functional MRI
Acoustic noise contamination is an unavoidable consequence of the
operation of current MRI devices. Its effects are complex, potentially
including stimulus masking, task difficulty modulation, interference with
selective attention and reduction in auditory system dynamic range.
37. Acoustic Noise Effects in Functional MRI
Subject motion relevant to imaging studies involving language and
communication can be categorized into three types: (1) rigid-body motion
of the head, (2) parenchymal motion of the brain, and (3) motion of the
articulators, particularly the jaw.
38. Motion Effects in Functional MRI
Subject motion relevant to imaging studies involving language
and communication can be categorized into three types: (1) rigid-
body motion of the head, (2) parenchymal motion of the brain,
and (3) motion of the articulators, particularly the jaw.
39. Why Require Overt Responses in Language
Experiments?
If the neural mechanisms of overt and covert speech responses are
sufficiently similar, or dissociable by subtractive analysis, it might be
possible to assuming that the neural correlates of overt speech production
are not unlike those of covert speech except for the associated motor and
auditory processes.
40. 4.4 Alternative Techniques for Functional MRI
One of these approaches is the so-called interleaved acquisition technique
which involves acquisition techniques in which task performance and image
acquisition are interleaved in time.
41. 4.5. Advantages and disadvantages of Interleaved
Acquisition Techniques
By employing a longer repetition time than customarily used in
conventional continuous imaging, interleaved acquisition techniques have
potential increased sensitivity to small signal changes, resulting from the
improved contrast-to-noise that occurs with longer TR intervals as a result
of more fully recovered longitudinal relaxation.
the proportion of time spent in performance of the target task is reduced,
resulting in a situation in which the subject makes relatively fewer responses
during the imaging session than during a continuous acquisition
experiment utilizing the same overall run length.