a) Recent concepts in theories of phonological development: Generative

phonology, natural
phonology, non-linear phonology, optimality theory

b) Application of phonological theories in evaluation and management of

phonological
disorders
• Children begin to vocalise from the moment they are born.
• Refinement of this vocalization into intelligible speech takes many
years while children’s body structures develop and their perception
and production systems become more sophisticated and attuned to
their ambient (native) language.
RELEVANCE OF UNDERSTANDING TYPICAL SPEECH SOUND ACQUISITION FOR SLPS

• In paediatric SLP practice, judgments about whether a child’s speech

is typical occurs daily.
• This decision making is guided by knowledge of research data as well
as clinicians’ experience.
• Studies of children’s typical speech sound acquisition are the primary
research data used by SLPs to make these decisions.
Seven main areas of SLP practice are informed by a comprehensive
understanding of speech sound acquisition:
1. Referral: Providing advice to parents, educators, and health professionals
regarding whether a child should be referred for a speech assessment.
2. Assessment: Deciding which assessment tools are appropriate to
examine speech behaviours that are relevant for the age of the child. For
example, if the child is 1.6, an inventory of consonants, vowels, and syllable
shapes should be determined. If the child is 7.0, polysyllabic words and
phonological awareness skills also should be assessed.
3. Analysis: Analysing the speech sample in order to decide whether the child’s speech is age
appropriate on a range of measures.
4. Diagnosis: Determining whether a child has a delay or disorder and whether his or her
areas of difficulty warrant speech sound intervention.
5. Selecting intervention targets: There are two major schools of thought regarding how to
use knowledge of typical speech sound acquisition to select intervention targets.
Proponents of the traditional developmental approach suggest that intervention targets should
focus on errors on the production of early developing sounds (cf. Davis, 2005; Shriberg and
Kwiatkowski, 1982b). Proponents of the complexity approach (also called nontraditional
approach and least knowledge approach) select later-developing sounds; the aim of such an
approach is to produce a systemwide change (cf. Gierut, 2007; Gierut, Morrisette, Hughes,
and Rowland, 1996)
6. Intervention: Adapting teaching and feedback to an age-appropriate
level, and determining that a child has achieved his or her goals to the
expected level.
7. Dismissal/discharge: Deciding whether a child’s speech is within
normal limits for his or her age (Tyler, 2005) and whether his or her
speech sound intervention should be concluded for other reasons, for
example, the child has progressed as far as he or she is going to and no
longer is making progress or the child lacks motivation for continued
therapy.
MODELS OF SPEECH ACQUISITION

• The acquisition of speech and language is a complex process whose precise nature is not known.
Many models of acquisition have been proposed. Each model provides a different insight into
how the process might work. Thus, it is useful for SLPs working with children to have
knowledge of a range of models.
• Barlow and Gierut (1999, p. 1482) recommend that when considering models of speech
acquisition, an adequate model must account for
(a) the actual facts of children’s productions and the mismatches between a child’s output and the
adult input forms;
(b) the generalities that span children’s sound systems, as well as associated variability within and
across developing systems; and
(c) the changes that occur in children’s grammars over time.”
• They also indicate that an adequate model must be “testable and falsifiable.”
• Traditionally, SLPs used a behaviourist model to explain how children learn the sounds of the
language.
• More recently, linguistic-based models have evolved based on theories of generative phonology,
natural phonology, nonlinear theory, optimality theory, and sonority. Such approaches not only
provided SLPs with insight into children’s developing speech systems but also extended and
enhanced guidelines for assessment, analysis, and intervention.
• Most linguistic theories maintain that innate or natural mechanisms govern a child’s phonological
system. These are expressed as distinctive features in generative phonology, phonological
processes (or patterns) in natural phonology, multitiered representations in nonlinear phonology,
and constraints in optimality theory.
• However, although linguistic-based models provided descriptions of children’s phonology, they
failed to provide explanations of the underlying cognitive mechanisms involved in the perception
and production of speech. Consequently, the most recent explanations are psycholinguistic models
of speech development.
Traditional Models of Speech Acquisition

Behaviourist Models
• Behaviourism focused on describing overt and observable behaviours.
• Behaviourism was proposed by Watson (1913/1994, p. 248) in his article “Psychology as a
Behaviourist Views It” and was described as an “objective experimental branch of natural
science that can be studied without references to consciousness.”
• The most influential proponent of behaviourism was B. F. Skinner. Skinner’s operant
conditioning can be traced back to the stimulus-response psychology of the Russian
physiologist Pavlov, who trained a dog to respond by salivating when hearing the stimulus of a
bell ringing.
• Skinner created the concept of operant or instrumental conditioning to focus on controlling
acts by changing the consequences that occur immediately following the act (Skinner, 1972;
Thomas, 2000).
• In a behaviourist approach, consequences can be described either as
positive or negative, reinforcement or punishment. Skinner’s
behaviourism has been applied to a wide range of ages, cultures, and
behaviours (including physical, social, and emotional).
• Behaviourist principles have been applied throughout SLP practice,
particularly from the 1950s to the 1970s
Application to Typically Developing Children:
• When behaviourist models have been applied to speech acquisition, the focus
has been on observing environmental conditions (stimuli) that co-occur and
predict overt verbal behaviours (responses).
• For example, Olmsted (1971) suggested that sounds that were easy to
discriminate would be learned first; however, his order has not been supported
by subsequent research.
• Behaviourist researchers documented normative behaviours of large groups of
children during the speech acquisition period. Speech-language pathologists
described speech development as correct pronunciation of speech sounds.
Drawback:
• A major criticism of the application of behaviourism to children’s speech
acquisition is that children master speech and language acquisition more
quickly than they could if they had to depend on stimulus-response
mechanisms to learn each element.
• That is, there is no capacity for parental/environmental reinforcement of all
speech behaviours, leading to mastery of such a complex skill as speech and
language.
• Another criticism is that acquisition of speech and language is too complex to
be explained solely by reinforcement.
Application to Speech-Language Pathology Practice:
• Although behaviourism as an explanation for sound acquisition has
not been supported, behavioural principles have had a significant
impact on SLP practice for children with speech sound disorders.
• Speech production for many years was considered a motor activity,
and analysis of speech was conducted using segmental error analysis.
Speech was analyzed as a series of sounds, and the function of sound
differences to signal meaning differences was not taken into
consideration (e.g., Van Riper and Irwin, 1958).
Intervention:
• The stimulus-response paradigm was the basis of traditional
articulation intervention. Using this behavioural approach, a child was
presented with a sound or word that he or she was required to say and
then received positive reinforcement in the form of praise, a sticker, or
a token on a set schedule (cf. Winitz, 1969)
Linguistic Models of Speech Acquisition

• In the traditional models of speech acquisition, there was limited

consideration of the patterns, structures, and contexts of
misarticulations (with exceptions such as McDonald [1964b]) or of the
cognitive dimension of linguistic knowledge. This gave rise to the
linguistic models of speech acquisition.
1. Generative Phonology

• Generative phonology is a theory of the sound structure of human languages and was
developed by Noam Chomsky. The term generative refers to the belief that speech sounds
are generated by transforming the underlying representation into a surface form using a
language-specific rule.
• The key principles of generative phonology were illustrated in the landmark study of English
phonology presented in Chomsky and Hallé’s (1968) book, The Sound Pattern of English.
• Generative phonology moved away from the traditional phonemic analysis and introduced
two major concepts:
1. Phonological rules map underlying representations onto surface pronunciations.
2. Phonological descriptions depend on information from other linguistic levels.
• The area of generative phonology that has received the most attention is the description of
phonological relationships that are expressed by proposing an abstract underlying
representation and a set of phonological rules.
• Generative phonology represents the application of principles of generative[or
transformational] grammar to phonology.
• Example : John is eager to please
John is easy to please
• If the two sentences are analyzed according to a structuralist point of view, the
results will indicate that both sentences have exactly the same structure.
• However, this analysis does not reveal that the two sentences have drastically
different meanings. In the first sentence, John wants to please someone (John is
the subject). In the second sentence, someone else is involved in pleasing (john
is object).
• One aim of generative grammar was to provide a way to analyze sentences that
would account for such differences.
• To do this, a concept was developed that postulated not only a surface level of
realization but also a deep level of representation.
• Competence and performance were also terms that distinguished between
surface and deep levels of representation.
• Language competence was viewed as the individual’s knowledge of rules of
language, whereas
• Performance was actual language use in real situations. Structuralists and
behaviorists focused on an individual’s performance.
• To provide an example, generative phonology can be used to explain the way that English-
speaking adults nasalize vowels before nasal consonants. In generative phonology, this concept is
written as the following rule:

• The information to the left of the arrow indicates the segments that conform to the rule. The
arrow means “is realized as.” Only the relevant rules are included to the right of the arrow. Other
features are assumed to remain as they were. The diagonal slash means “in the context of.” The
dash and information that follow provide the context of the segment described by the rule.
• Thus, this generative phonology rule reads: Vowels are realized as nasal in the context of (in this
case, specifically just before) nasal consonants.
Application to Typically Developing Children:
• Generative phonology has been applied to the understanding of children’s speech acquisition
(cf. Grunwell, 1987) as it enabled description of the relationship of children’s productions to
adult pronunciation in terms of phonological rules.
• Grunwell indicated that generative phonology has been readily applied to children’s speech
because generative phonological rules can explain substitutions, distortions, omissions,
additions, metathesis, and coalescence.
Drawbacks:
• Some of the premises of generative phonology have received criticism in subsequent research,
there has been criticism of the premise that the child’s underlying representation of the sound is
adultlike.
• Additionally, there has been criticism of the premise that the rules that were applied had a
corresponding reality to the processing and production systems of the child (i.e., it is not clear
that we actually apply such rules in our heads when we comprehend and produce speech).
Application to Speech-Language Pathology Practice:
• IN ASSESSMENT: Comparison of child’s phonological system to the
adult’s – using the distinctive features, Description of the patterns of errors –
using the distinctive feature analysis, Several sound substitution can be
compared to the target phoneme, Identification of SODA errors.
• IN MANAGEMENT: Compares the phonetic features of target sound with
the substituted sound – useful in therapy, captures those phonetic features
that distinguish between the phonemes of a language, Distinctive Feature
analysis is also used in Feature Geometry, Distinctive feature approach is
based on this.
2. Natural Phonology
• The theory of natural phonology (Stampe, 1969, 1979) formed the basis of the phonological process approach
to assessment and treatment of speech sound disorders and is regarded as the phonological model that has had
the greatest impact on the field of SLP (Edwards, 2007).
• Natural processes (or patterns) are those that are preferred or frequently used in phonological systems and are
identified in two ways:
(1) those that are universal across languages and
(2) those that are frequently used by young children.
• According to Stampe, a phonological process is a “mental operation that applies in speech to substitute for a
class of sounds or sound sequences presenting a common difficulty to the speech capacity of the individual, an
alternative class identical but lacking the difficult property” (1979, p. 1), and phonological processes merge “a
potential opposition into that member of the opposition which least tries the restrictions of the human speech
capacity” (1969, p. 443).
• In Stampe’s view, the child’s underlying representations are akin to adult forms. Natural (or innate)
phonological processes apply to these underlying representations, resulting in the child’s productions (or
surface forms).
• For example, it is assumed that children have the adult form of a word, such as tree /tri/, in their underlying
representation. However, natural processes such as cluster reduction are applied because the child (at least
temporarily) has some limitation to produce a particular sound or group of sounds. In this case, the surface
form (child’s production) would most likely be [ti].
Drawbacks:
• A shortcoming of this theory is that some errors may fit into more than one category. For example,
if a child attempted to say dance /dæns/ and said [dæn] instead, it is not clear if this is an example
of final consonant deletion, cluster reduction, stridency deletion, or some combination of these.
Application to Typically Developing Children:
• Natural phonology has provided insight to the understanding of typical speech acquisition.
• Natural processes are described as innate rules that are systematically applied to speech production
until children learn to suppress them. Because these rules are universal, they are meant to apply to
all children speaking all languages.
• Thus, speech acquisition is a progression from these innate speech patterns to the pronunciation
system of the language(s) learned by the child.
• By applying natural phonology to English speech acquisition, Grunwell (1987) presented a table
of the ages of suppression of phonological processes by typically developing children, such as
cluster reduction, fronting, and stopping.
• Other researchers have also provided lists of natural phonological processes (e.g., Ingram,
1976; Shriberg and Kwiatkowski, 1980). Shriberg and Kwiatkowski advocated the clinical
use of eight “natural processes”: (1) final consonant deletion, (2) velar fronting, (3)
stopping, (4) palatal fronting, (5) liquid simplification, (6) cluster reduction, (7)
assimilation, and (8) unstressed-syllable deletion.
Application to Speech-Language Pathology Practice:
• The phonological pattern/process approach to assessment and intervention based on natural
phonology transformed the way that SLPs viewed children’s speech sound errors.
• Since Ingram’s (1976, 1989a) seminal work on phonological impairments in children, SLPs
increasingly have applied descriptive linguistic-based models to their clinical activities.
• Ingram’s application of natural phonology was widely accepted by SLPs in the 1970s and
1980s and remains popular for directing the assessment, analysis, and intervention of
children with speech sound disorders (Bankson and Bernthal, 1990a; Khan, 1982; Shriberg
and Kwiatkowski, 1980; Weiner, 1979).
• Assessment approaches were developed to specifically assess subgroups of
sounds within a given phonological pattern (e.g., Bankson-Bernthal Test of
Phonology [BBTOP] [Bankson and Bernthal, 1990a]).
• Phonological processes were also described as part of a broader analysis
procedure for several speech-sampling tools (e.g., Phonological Assessment of
Child Speech [PACS] [Grunwell, 1985]) and as stand-alone analyses to be
applied to conversational speech (e.g., Natural Process Analysis [Shriberg and
Kwiatkowski, 1980]).
• One of the goals of intervention based on natural phonology is “to teach
children to suppress innate simplification processes” (Hodson, 2010b, p. 55).
Limitations :
• First, although most SLPs can readily describe children’s nonadult productions using
phonological process terms such as cluster reduction and fronting, SLPs’ use of phonological
processes are descriptive rather than an application of the theoretical tenets of natural phonology.
• Shriberg (1991, p. 270) described this as an “atheoretical use of process terminology.”
• Second, natural phonology does not account for “nonnatural” simplifications in children’s speech
(Hodson, 2010a). Many children with highly unintelligible speech produce speech sounds in a
way that cannot be classified using natural phonology. Terms such as backing and initial
consonant deletion are in the literature to describe phonological processes that are not seen in
children with typical speech acquisition (Dodd, 1995b).
• One question that remains unresolved with natural phonology is whether the process labels being
applied actually represent mental operations going on inside the head of the child.
• However, because such labels do capture “patterns” of errors being observed, the term
phonological patterns is frequently used in place of phonological processes. For example, the title
of a popular assessment tool in this area is the Hodson Assessment of Phonological Patterns
3. Nonlinear Phonology

• Nonlinear phonology refers to a collection of theories that focus on the hierarchical nature of the
relationships between phonological units.
• Goldsmith introduced nonlinear phonology in his doctoral dissertation (1979) and later expanded
upon it (Goldsmith, 1990). These theories include auto segmental theory, metrical theory, moraic
theory, feature geometry theory, and under specification theory.
• Nonlinear phonology attempts to account for the idea that production of speech involves more than
just production of a sequence of phonemes; it takes into account many elements (features, segments,
syllables, feet, words, and phrases) both independently and in relation to one another; hence, the
term nonlinear.
• There are two main tiers in nonlinear phonology:
1. The prosodic tier focuses on words and the structure of words and includes a number of levels:
word tier, foot tier, syllable tier, onset-rime tier, skeletal tier, and segmental tier (see Figure 3.1).
2. The segmental tier focuses on the segments or speech sounds and the features that make up those
sounds (see Figure 3.2).
• In the prosodic tier, the word tier simply denotes words.
• Immediately below the word tier is the foot tier, which refers to grouping of syllables, and syllables may be
either strong (S) or weak (w).
• A foot can contain only one strong syllable (but can also contain other weak syllables). A foot that includes
a weak syllable can be either Sw (left prominent, or trochaic), or wS (right prominent, or iambic).
• Below the foot is the syllable tier. A syllable consists of one prominent phoneme (the peak), which is
usually a vowel and less prominent phonemes (generally consonants) that can appear before or after the
peak.
• Consonants that appear before the vowel are known as onsets, and consonants that appear after the vowel
are codas. The peak and the coda together make up the rime.
• All languages allow syllables without a coda, which are sometimes called open syllables (e.g., CV). Some
languages do not allow for closed syllables that have a coda (e.g., CVC). Across the world’s languages,
open syllables occur more often than closed.
• Below the syllable tier is the skeletal tier, which includes slots for the individual speech sounds
• In the segmental tier, features are described according to three nodes: the root node, the laryngeal
node, and the place node (see Figure 3.2).
• The root node [sonorant] and [consonantal] defines the segment as a vowel/glide or a consonant.
• The features [continuant] and [nasal] define the classes of stops, fricatives, and nasals.
• The laryngeal node includes the features of [voice] and [spread glottis] and differentiate vowels
as well as voiced from voiceless consonants.
• The place node designates the oral cavity characteristics of the segment and includes labial
[round], coronal [anterior], [distributed], dorsal [high], [low], and [back].
• Default nodes are generally the most frequent features (unmarked) and the easiest features for a
child to use. In English, the default consonant is /t/ because it is coronal but not continuant, not
lateral, not nasal, and not voiced.
Application to Typically Developing Children:
• Two of the major benefits of nonlinear phonology to the understanding of typical phonological development
are
(1) the concept of links between the segmental and suprasegmental tiers and prosodic variables that highlight
the interaction between speech sounds and other speech-language domains and
(2) the view that development is progressive or additive, which is in contrast to the negative progression
suggested by a phonological process approach whereby children learn to “undelete” deleted final consonants
(Bernhardt and Stoel-Gammon, 1994).
• Bernhardt (1992a) discussed these developmental implications of nonlinear phonology and highlighted the
importance of considering children’s representations instead of the negative progression of rules of
generative grammar.
• Bernhardt suggested that children’s representations (or simple syllable templates such as CV units and stop
consonants) account for a large proportion of young children’s speech production and that children add to
these representations as they mature.
• Nonlinear phonology also enables consideration beyond the acquisition of consonants, as it also addresses
vowels, syllable shapes, words, and stress.
Application to Speech-Language Pathology Practice:
• Bernhardt and Stoel-Gammon (1994) presented an excellent overview of the benefits of
nonlinear phonology to SLPs.
• Bernhardt and Stemberger developed a theoretical text (1998), a workbook for clinicians
(2000), and a computerized assessment and analysis (CAPES: Computerized Articulation
and Phonology Evaluation System) (Masterson and Bernhardt, 2001).
• Additionally, Bernhardt and Stemberger have applied principles of nonlinear phonology to
goal setting and intervention for children with speech sound disorders (Bernhardt, 1992b;
Bernhardt et al., 2010).
• In the assessment of children’s speech using nonlinear phonology, attention is paid to
production of consonants as well as vowels, syllables, word shapes, and stress patterns.
• Nonlinear intervention goals focus on utilizing established sounds in new syllable shapes
and new sounds in established syllable shapes
4. Optimality Theory

• Optimality theory was first described by Prince and Smolensky (1993) in their report Optimality
Theory: Constraint Interaction in Generative Grammar. Optimality theory originally was developed
to describe adult languages. Its basic units are constraints, which are of two major types:
1. Markedness constraints (also called output constraints) capture limitations on what can be
produced (or the output). Output is simplified by markedness constraints that are motivated by the
frequency and distribution of sounds in the ambient language as well as perceptual and articulatory
characteristics of the sounds. Sounds that are difficult to pronounce or perceive are marked.
2. Faithfulness constraints capture the features to be “preserved,” prohibiting addition and deletion
that violate the ambient language.
• Constraints are assumed to be universal to all languages (Barlow and Gierut, 1999). There is a
reciprocal relationship between faithfulness and markedness constraints (Kager, 1999).
Faithfulness can result in the inclusion of phonological features, whereas markedness can result in
their exclusion.
Application to Typically Developing Children and Speech-Language Pathology Practice:
• Optimality theory has been applied to the understanding of the typical development of children’s speech
(Barlow and Gierut [1999]).
• The aim during development is for the output to match the adult target, which is achieved by promoting
faithfulness constraints and demoting markedness constraints.
• One example of the application of optimality theory has been to develop a model of children’s acquisition of
polysyllabic words (James, van Doorn, and McLeod, 2008). The model proposes five stages, each of which
elucidates relevant faithfulness constraints.
In stage 1 (ages 1;0 to 2;3), children are faithful to the stressed syllable and the duration of the whole word.
Stage 2 (ages 2;4 to 3;11) focuses on faithfulness to the number of syllables in a word. That is, during this
stage, children reduce the occurrence of weak syllable deletion, cluster reduction, and final consonant deletion.
Stage 3 (ages 4;0 to 6;11) highlights faithfulness to all the phonemes in the words and can co-occur with a
period of dysprosody (e.g., word rhythm is disrupted).
Stage 4 (ages 7;0 to 10;11) emphasizes faithfulness to word rhythm with accurate delineation between stressed
and unstressed syllables.
The final stage (age 11;0 and older) represents adultlike production of polysyllabic words.
• Optimality theory also has been applied to SLP practice with children with
speech sound disorders (Barlow and Gierut, 1999).
• Children’s possible productions of words are placed into a constraint table,
and violations of constraints are identified. Fatal violations (involving a
highly ranked constraint) are also identified.
• The optimal constraint (the form most likely to be produced by the child) is
the one that violates the least number of constraints, or the lowest-ranking
constraints.
• As children’s speech systems mature, they rerank constraints until eventually
their productions match the adult form (Edwards, 2007)
Studies

• Speech sound disorder at 4 years: prevalence, comorbidities, and

predictors in a community cohort of children
Patricia Eadie, Angela Morgan ,Obioha C Ukoumunne ,
Kyriaki Ttofari Eecen ,Melissa Wake ,Sheena Reilly
Aim
• The epidemiology of preschool speech sound disorder is poorly understood. Our
aims were to determine: the prevalence of idiopathic speech sound disorder; the
comorbidity of speech sound disorder with language and pre‐literacy difficulties;
and the factors contributing to speech outcome at 4 years.
Method
• One thousand four hundred and ninety‐four participants from an Australian longitudinal cohort completed
speech, language, and pre‐literacy assessments at 4 years. Prevalence of speech sound disorder (SSD) was
defined by standard score performance of ≤79 on a speech assessment. Logistic regression examined
predictors of SSD within four domains: child and family; parent ‐reported speech; cognitive ‐linguistic; and
parent‐reported motor skills.
Results
• At 4 years the prevalence of speech disorder in an Australian cohort was 3.4%. Comorbidity with SSD was
40.8% for language disorder and 20.8% for poor pre ‐literacy skills. Sex, maternal vocabulary, socio ‐economic
status, and family history of speech and language difficulties predicted SSD, as did 2 ‐year speech, language,
and motor skills. Together these variables provided good discrimination of SSD (area under the curve=0.78).
Interpretation
• This is the first epidemiological study to demonstrate prevalence of SSD at 4 years of age that was consistent
with previous clinical studies. Early detection of SSD at 4 years should focus on family variables and speech,
language, and motor skills measured at 2 years.
• Phonological/Traditional Approaches to Articulation Therapy
• A Retrospective Group Comparison
• Author:Edward S. Klein
• Method: A retrospective design was used to compare the relative
efficacy of traditional and phonological approaches in the treatment of
children with multiple articulation disorders. Nineteen children
received evaluation and therapy based on a traditional paradigm and
17 children received phonologically based evaluation and treatment
procedures.
• Results: indicated that children in the phonological group showed
significantly greater improvement in a significantly shorter period of
time than children in the traditional group.
• Nonlinear Phonology :Introduction and Clinical Application
Barbara Bernhardt and Carol Stoel-Gammon
• The purpose of this tutorial is to introduce a recent advance in phonological theory,
“nonlinear phonology,” which differs fundamentally from previous theories by focusing on
the hierarchical nature of relationships among phonological units. We first introduce the
basic concepts and assumptions of nonlinear phonological theory and then demonstrate
clinical applications of the theory for assessment and intervention. Data from a child with a
severe phonological disorder are used to illustrate aspects of nonlinear theory. The data are
first analyzed in terms of phonological processes in order to provide the readers with a
familiar starting point for comprehension and comparison. The nonlinear frameworks are
shown to provide a deeper analysis of the child’s phonological system than the phonological
process analyses and to lead to a more clearly defined intervention plan.
References:
• Bernthal, J. E. (2009). Articulation and phonological disorders. Speech
Sound Disorders in Children.
THANK YOU