Young clients with speech-sound deviations comprise the largest group of children receiving speech-language services (see Table A [PDF] online; ASHA, 2010). Practicing speech-language pathologists often use the term "articulation" to refer to mild or moderate speech-sound disorders and either "phonology" or "apraxia" to refer to severe or profound issues. The umbrella term preferred by ASHA, however, is "speech-sound disorders" (Bernthal, Bankson, & Flipsen, 2009).

Major Early Contributions  

The earliest published comprehensive method for working with children with speech-sound disorders was developed by Charles Van Riper (1939). Although there have been many adaptations (e.g., behaviorism overlay) from his original work more than 70 years ago, it has survived the test of time and is still the foundation for the practice of most SLPs. This phoneme-oriented approach is considered adequate for children with mild speech-sound disorders (e.g., lisp), but has been found to require an incredible amount of time for children with highly unintelligible speech as they "perfect" one sound (or cognate pair) at a time. SLPs have reported that phoneme-oriented methods work well for a child who has two or three sounds in error, but a child who has a dozen or so missing sounds may take five or six years for each sound to reach a criterion (e.g., 90%) in each position of words, in phrases, in sentences, and in conversation.

The landmark book, Phonological Disability in Children, by David Ingram (1976), provided a bridge between linguistic phonology and speech-sound disorders. His work helped SLPs look beyond individual phonemes to patterns of deviations, referred to as phonological "processes," and paved the way for new phonologically based remediation principles and procedures as well as phonological assessment instruments.

Intelligibility Considerations 

Bishop and Adams (1990) posited in their critical age hypothesis that children need to be intelligible by 5½ years of age or they are likely to have difficulty with decoding and spelling. Children with severe/profound speech-sound difficulties frequently demonstrate poorer phonological awareness skills than their typically developing peers, which, in turn, negatively affect literacy acquisition (Gillon, 2004). There is some urgency, therefore, for children to be intelligible before entering school.

The method generally recommended for assessing intelligibility is to obtain a connected-speech sample of more than 100 words and have at least one unfamiliar listener write down words that can be understood, placing a dash for words that cannot be identified. The number of words understood is divided by the total number of words in the sample to obtain an intelligibility percentage.

A guideline for expected intelligibility (originally proposed by Caplan and Gleason, 1988) can be calculated by dividing the child's age in years by four and converting that number into a percentage:

• 2-year-old: 50%
• 3-year-old: 75%
• 4-year-old: 100%

Thus a 4-year-old who is 50% intelligible is considered to be "delayed" by two years. Intelligibility is a much more meaningful measure for young children than are phoneme acquisition normative data, which vary greatly from study to study (see Smit, 1986, for a review of some of the issues related to phoneme acquisition norms). Because different school districts in the United States use different sets of norms, a child may qualify for speech-language services in one school district but may not qualify in another.

Expediting Intelligibility Gains 

In 1975, we (graduate students and author) began an experimental phonology clinic at the University of Illinois (see Hodson, 2010) that accepted only children with highly unintelligible speech (intelligibility generally between 0% and 15% in spontaneous speech samples at the onset of treatment). In the first three years, as hypotheses were formulated and tested, we made drastic changes. For example, it soon became apparent that targeting word-initial singleton /f/ or /s/ first was a mistake for children who substituted stops for stridents and also reduced /s/ clusters. As we experimented with targeting /s/ clusters before singleton stridents for these children, they achieved major gains in intelligibility. Moreover, most of the children generalized to singleton stridents, eliminating the need to target them. During this period, we explored the concept of focusing on phonological patterns (e.g., final consonants, /s/ clusters), with phonemes serving as a means to an end rather than the end goal (e.g., 90% mastery of phonemes, one at a time).

We also found that the emphasis on "counting and charting errors" was counterproductive for these young clients with highly unintelligible speech. Instead, our goal was for the child to produce the specified target pattern 100% correctly in carefully selected production-practice words, incorporating "assists" (e.g., tactile cues, amplification) as needed, and phasing out the assists as the child gained facility in producing the target. Every time children continued making the error in their production-practice words, they were reinforcing the inaccurate kinesthetic image. Thus, it is critical that the child be "stimulable" and capable of producing the target sound (usually with assistance at first) in order to develop the new accurate kinesthetic image. We would, of course, "stimulate" nonstimulable sounds (e.g., /k/) for a few minutes during each session (i.e., teach stimulability) and then target these as soon as the child could actually produce them.

We also explored targeting phonological patterns via cycles (time periods varying from five to 16 hours, depending on the number of patterns that needed to be targeted and also stimulability). Typically a phoneme (or consonant cluster) is targeted one hour per week (i.e., one 60-minute session, two 30-minute sessions, or three 20-minute sessions), with each pattern typically being targeted from two to five hours per cycle. We also realized that phonological patterns needed to be divided into primary (those targeted first and then recycled as needed until they began emerging in conversational speech) and secondary patterns (see Tables 1 [PDF] and 2 [PDF]). It is important to note that complexity is increased gradually so that the client is optimally challenged but successful from the beginning of treatment. Most young children in our phonology clinics (Wichita State University, San Diego State University, University of Illinois) were judged to be essentially intelligible within three to four cycles (i.e., approximately 30 to 40 contact hours) and simultaneously demonstrated vastly improved phonological systems.

Underlying Concepts, Theoretical Underpinnings 

The approach that has emerged from this research—the cycles approach to phonological remediation—is based on developmental phonology theories and cognitive psychology principles as well as on ongoing clinical phonology research. The cycles approach most closely aligns with the gestural phonology theory (Browman & Goldstein, 1986). A basic tenet of gestural phonology is that phonological representation is based on speech perception as well as on "speech production physical constraints." Seven underlying concepts serve as the basis for the cycles approach (Hodson, 2007; Hodson & Paden, 1991).

• Phonological acquisition is a gradual process (Ingram, 1976). Children typically do not learn each sound to a 90% criterion and then learn the next sound to a criterion. There is considerable vacillation and experimentation on the part of typically developing young children (Dyson, 1988). This principle is the major reason for the cycles approach, which enhances phonological patterns in a sequential manner, and then recycles patterns as needed until the patterns begin carrying over into conversation.
• Children with normal hearing typically acquire the adult sound system primarily by listening (Van Riper, 1939). Most parents do not tell children where to place their tongues. Children need optimal hearing and adequate perceptual skills during their early speech-learning years (Shiller, Rvachew, & Brosseau-Lapre, 2010).
• Children associate kinesthetic and auditory sensations as they acquire new phonological patterns, enabling later self-monitoring (Fairbanks, 1954). We incorporate production practice to help children learn new accurate kinesthetic images. When we imitate children's error productions, they usually tell us that we are wrong—but then repeat their error productions because they rely on inaccurate kinesthetic images and fail to perceive their own errors. We also incorporate slight amplification during a listening activity at the beginning and end of each session to help children learn to perceive differences between errors and desired productions.
• Phonetic environment can facilitate (or inhibit) correct sound productions (Kent, 1982). Production-practice words must be chosen with extreme care during beginning cycles. Words with phonemes at the same place of articulation as the error should be avoided until later cycles (e.g., "cat" is inappropriate at first for a child who demonstrates velar fronting because of alveolar assimilation effects).
• Children tend to generalize new speech production skills to other targets (McReynolds & Bennett, 1972). We know that children generalize (and sometimes overgeneralize). The key is to target optimal patterns to facilitate generalization and increase intelligibility and then proceed on to other patterns, with a period of rest for a pattern before recycling it.
• An optimal match facilitates a child's learning (Hunt, 1961). Phonological analysis determines the child's match (e.g., Hodson, 2003, 2004) so that treatment can begin one "step" above the child's current functioning level, resulting in the child being optimally challenged but also successful from the beginning of treatment.
• Children are actively involved in their phonological acquisition. This principle is important for children with speech-sound disorders as well as for typically developing children. Young children need to be actively engaged participants rather than passive imitators.

Phonological Remediation 

The structure described below has been used successfully in schools, hospitals, clinics, and private practice.

• The child reviews production-practice picture cards from the previous session. These cards are then set aside (and may be used again along with new, more complex words during a later cycle, depending on whether the phonological pattern needs to be recycled). The SLP then reads a list of approximately 20 words that contain the new target pattern for the week. The child is to listen attentively, but must not repeat these words. Slight amplification is provided during this 30-second listening activity.
• The child participates in experiential-play production-practice motivational activities (e.g., bowling, flashlight game), naming pictures and objects of four or five carefully selected target words with the week's pattern before taking a turn in the activity. The SLP changes activities every eight to 10 minutes. Many of these activities are repeated during ensuing weeks.
• A metaphonology activity is incorporated (e.g., rhyming, segmentation) to enhance the child's phonological awareness skills.
• The SLP probes for the optimal phoneme target for the next week (within the phonological pattern designated from phonological assessment results). For example, if the current pattern is /s/ clusters, the clinician models words with various /s/ clusters (e.g., spot, store, snow). The cluster that the child produces most successfully becomes the target for the ensuing week.
• The listening list from the beginning of the session is read to the child again with slight amplification.
• Parents receive the listening list and the production-practice picture cards and are asked to provide two minutes of home practice every day.

Adaptations for Toddlers 

Most of the 2-year-olds referred to our clinic are not willing to participate in regular production-practice activities. Some are nonverbal; others are unwilling to name pictures or imitate words. These children participate in a cycle (typically a semester) of focused auditory input/stimulation for the primary patterns.

The SLP fills the room with objects and activities for a primary pattern phoneme. For example, objects and tasks for final /p/ (e.g., mop, top, tape, rope, hop, up, cup, nap, pup, peep) are incorporated to stimulate awareness of final consonants. The child participates in the play activities but is not asked to name words during this cycle. The parents receive a list of words (with the week's target pattern) to read to the child each night, but they do not ask the child to say words at this time. These adaptations provide a great foundation and the child readily moves to production practice during the next cycle.

Evidence Considerations and Research Needs 

There is consensus among SLPs that treatments for speech-sound disorders are successful. Baker and McLeod (2010) used the levels-of-evidence system selected by ASHA (adapted from the Scottish Intercollegiate Guideline Network) to review 134 studies. One of their observations was that comparative studies are generally lacking. A great need exists for a large, well-designed controlled experimental study comparing outcomes of various interventions for children with highly unintelligible speech. The number of participants in each treatment group needs to be large because of the heterogeneity of clients with speech-sound disorders (e.g., Harbers, Paden, & Halle, 1999). Ideally, this study should be conducted by investigators independent of the individuals who created these approaches to minimize actual bias or even the appearance of bias. It is critical, however, that the approach creators be consultants for purposes of fidelity.

There also is a need for more advocacy so that treatment protocols that optimize outcomes—rather than top-down factors such as scheduling issues—determine intervention approaches.