Teaching – Week 26 – Phonological Approaches

Teaching – Week 26 – Phonological Approaches to Intervention

The last chapter of Developmental Phonological Disorders and the last week of the course is devoted to phonological approaches to intervention, specifically covering (1) core vocabulary approach; (2) cycles approach; (3) minimal pairs procedure; (4) nonlinear approach; (5) metaphon; and (6) classroom curricula for emergent literacy. Most children with developmental disorders will almost certainly require a phonological approach to intervention at some point during their intervention program. Furthermore some studies have shown that phonological approaches are more efficient than a traditional approach to therapy. Somewhat surprisingly, given Barbara Hodson’s heroic efforts over the past 30 years to educate the profession about the importance of a phonological approach and her particular application (the “cycles” approach) a new survey by Brumbaugh and Smit (2013) found that the majority of SLPs in the United States reported a preference for the traditional approach although just over fifty percent also used the cycles approach and/or minimal pairs procedures often or sometimes. I was surprised to see that 19% of respondents reported that they were unfamiliar with the cycles approach but not surprised that 75% were unfamiliar with nonlinear phonology. Overall the authors noticed that the responding SLPs tended to be unfamiliar with approaches and procedures that were introduced to speech-language pathology more recently than about 1985, regardless of how recently the responding SLP had graduated.

One of the more interesting findings of the Brumbaugh and Smit’s survey is that only 3% of respondents reported being unfamiliar with nonspeech oral motor exercises (NSOMEs) but only 28% used them often or sometimes. This was an unusual pattern as more typically low use of an approach was associated with low familiarity. For example, we cannot fault SLPs for not using a nonlinear approach when so few know anything about the requisite assessment or target selection procedures. This is an important gap that we are trying to close with our book. There are some children with complex phonological patterns that really cannot be adequately treated without knowledge of multilinear (i.e., nonlinear) phonology. We present two such cases in the book, one in Chapter 11. Some of the data from the Chapter 11 case  is attached here: Rvachew DPD Demonstration 11-2. This child was initially treated with the cycles approach but he did not make good progress and was refusing to cooperate with therapy or home practice activities. These are the junctures at which an SLP might resort to a NSOMEs even knowing that there is no evidence for their effectiveness – the child will enjoy them, it looks like progress is being made because they are easy and the child will experience success for a change and maybe the SLP will stumble upon a better idea or the child will mature past the stumbling block while the program plays itself out. However, in this specific case a multilinear analysis of the child’s speech revealed the heart of the problem in the prosodic tiers of the phonological hierarchy, leading to a successful reorientation of the treatment program. So, one part of the battle against inappropriate use of NSOMEs has been all but won – awareness has spread to almost the entire profession and these procedures are not used when the SLP has a better option. The other part of the challenge is to ensure that SLPs have the foundational knowledge and clinical skills to apply the better options for all cases, including complex children such as the young boy described in the attached case study.

Weeks 24 and 25 – Output Oriented Approaches to Intervention

Even when devoting 2 weeks to Chapter 10 of Developmental Phonological Disorders: Foundations of Clinical Practice, there is too much material in the chapter to cover it all in class. I describe in detail many procedures for improving articulation skills in this chapter, linking the procedures to different stages of speech development as described in Chapter 3 on the development of speech motor control. Specifically, the material covered in Chapter 10 is as follows: (1) Vocal play, linked to the expansion stage in early infancy when the infant produces increasingly variable vocal outputs; (2) two different but similar sensory motor approaches to speech therapy involving practice with nonsense syllable drills; this section is linked to the canonical babbling stage when the infant learns to coordinate respiration, phonation and articulation to produce speech-like syllables and early words; (3) the traditional (Van Riperian) approach to speech therapy to teach new and more complex segments and syllable structures; this section is tied to early speech development when the child achieves intelligible speech through system expansion; and finally (4) application of the principles of motor learning, including the importance of self-monitoring, to promote refinements to the child’s articulatory precision as speech and language demands increase throughout the school years. As I said, the depth and breadth of material in this chapter is too great to cover in 6 hours of class, especially since the students spend at least two of those hours demonstrating phonetic placement techniques to each other. I focus largely on the traditional approach to therapy because there are many children for whom these techniques are still relevant. Furthermore, the students draw their cases for their term projects from the school population meaning that quite a few of these cases involve residual errors in the production of /s/ or /ɹ/. I also make time to lecture on the application of the sensory motor approaches for apraxia of speech because this used to be my clinical specialty so it is my favorite topic.

Ultimately I don’t have time to spend covering techniques for encouraging vocal play in detail which is unfortunate. Sometimes during their spring practicum the students email with frantic requests for information when they are required to treat young children or children with very restricted phonetic repertoires who do not respond to imitative approaches to speech therapy. What should they do? Follow the vocal play approach described in Chapter 10 of course. I never know whether to be annoyed or amused when they tell me that they didn’t pack their book because they didn’t have room in their luggage! Fortunately the material comes directly from this publication, Dethorne et al. (2009), so I can direct them to an accessible source. However, the theoretical justification, basic science background, and demonstrations on how to apply the technique can only be found in my book so I really wish they would make room for it in their suitcases. Those students who do, which fortunately is most of them, find it very helpful when they are explaining to their clinical educators what they are doing.

Week 22 – Input Oriented Approaches to Intervention

In the final three weeks of the winter term we discuss clinical procedures for treating phonological disorders. The approach taken is to expect the students to read about the approaches and procedures in the book while I meet with the students individually to help them decide how to choose and apply procedures to the case that they have been working through for their term project. I also show videos and discuss case examples in class. The videos and case examples come from my research projects. The specific therapeutic procedures that are described in Chapter 9 are (1) focused stimulation; (2) ear training, which includes four procedures, identification, isolation, stimulation (i.e., auditory bombardment), and discrimination (i.e., error detection); and (3) dialogic reading. These procedures serve to provide rich and varied language input to the child, strengthening their acoustic-phonetic representations for words, which in turn form an essential foundation for the development of articulatory-phonetic and phonological representations.  I have attached a table ( Input Oriented Approach Example Table) that describes the input oriented procedures that we used successfully in a recent randomized controlled trial (along with minimal pairs procedures, described in Chapter 11 of our book – i.e., the chapter on  phonological approaches). I will be publishing the results of a randomized control trial involving this approach next year: we showed an advantage for speech accuracy and phonological awareness skills after 12 weeks when using these procedures in combination with parent training in dialogic reading and a small group phonological awareness intervention.

During the summer months I have the pleasure of observing how students apply this learning in the clinic in the context of my clinical trials. Generally they do a terrific job of learning to apply these procedures and my studies show that they are very effective in the treatment of phonological disorders even though the student clinicians often apply the procedures imperfectly during the first few weeks. The text describes the application of the procedures in some detail and the procedure manuals for my studies provide very explicit examples of activities that can be used but even with these supports it takes a while for students to get it exactly right.

In particular students have some difficulty devising error detection tasks that are meaningful. For example, I describe for the students an error detection task in which the child controls a toy dog and the student SLP calls the dog for dinner; let’s say the target is /θ/ and the dog’s name is Thor. The idea is that if the student SLP calls “Thor” the dog should come running for dinner.  If the student SLP misarticulates the dog’s name (e.g., “For”) the dog should lie down in bed. The child makes the judgment about whether the student SLP has called the dog’s name and manipulates the toy dog accordingly. To my mind, the responses associated with the correct and incorrect articulations of the dog’s name are meaningful and sensibly related to the game. The students come up with their own activities as new words are added to the practice set and often the game and the response set don’t make sense at all (I won’t even describe the wackier ones). One popular game was not too bad: the student SLP would produce a set of words with /θ/ onsets; the child’s job was to feed a bone to the dog when the word was correctly articulated; alternatively the child was supposed to put the bone in the garbage if the word was incorrectly articulated. Not a bad activity, it’s fun and there are clearly differentiated responses for correct and incorrect articulations of the word; but ultimately it doesn’t make a lot of inherent sense and sometimes the children have trouble figuring out what they are supposed to do. It can be a bit confusing: the clinician says something like “thistle” and the child is supposed to give a bone to “Thor” and the clinicians says “forn” and the child is supposed to put a bone in the garbage. The students see the game as being analogous to the SAILS clinical software in which the child points to the “thin” picture if they hear the word “thin” and to an X if the word is “not thin”. But it is really not the same task at all. The meaning relation between the stimulus and the response has been broken.Thor activity

I am thinking that I may have made a mistake in using Van Riper’s original terminology when I described his approach in Chapter 9 (he called this type of task a “discrimination” task but as you read his description of how to implement it is clear that he is talking about strengthening the child’s underlying perceptual representations for phonological categories rather than phonetic discrimination). I still think that his ear training activities are valuable (although SAILS with variable voice input is more effective). But it is not quite right to describe step 4 of the program as discrimination. I view speech perception as a skill that is fundamentally about deriving meaning from sound rather than about hearing “differences” between speech sounds. It is important to design meaningful activities for children with phonological disorders.

When talking to SLPs I am finding that Chapter 9 is a section of the book that stimulates them to say things like “I wish I had known about these techniques before!” even though the procedures are not new at all – it is simply that they are not currently applied in the context of speech sound disorders. It is as if we are forgetting (again) that the speech sound system cannot be disconnected from phonology, and furthermore, phonology is all about language. These procedures are good options for children who are not responding well to the traditional approach and especially for children who are too young or shy to cooperate well with structured articulation practice. They are especially important for children who have imprecise phonological representations (see Ramus et al. 2013).  However, it takes students a while to learn to implement the ear training procedures exactly right because there are so few models of correct implementation in the clinical community, given their long disuse . I hope that this will change over time.

Week 21 – Goal Selection

Continuing with the theme of expecting the students to read the material before the class and using class time to practice skills, Chapter 8 of Developmental Phonological Disorders provides several demonstrations of how to select targets (following from the case studies in Chapter 5 on assessment and Chapter 6 on speech sample analysis) and a new speech sample to use for practice. I have attached the sample here (Rvachew DPD Case Study 8-5) for anyone who may want to use it as an in-class or laboratory activity although samples that you have collected are always better because you can provide the background on the child in a more personal way. Working through this sample in class occurred just before the students’ midterm take-home exam (involving another sample from an actual child) but was also intended as practice for their term assignment (in which they had to design a treatment program for a child they had assessed). Their performance on both of these assignments was impressive this year so I know that the book and the in-class practice prepared them well for this difficult task.

When working through the options for treatment targets for Case Study 8-5, we used the framework described by Pamela Grunwell, but with our choices informed by the principles of goal selection from a nonlinear perspective (Table 8-4 of the book). I really like this approach – there is no silliness about selecting complex or simple targets! You end up with a very nice mix every time. Here is the framework suggested by Grunwell:

  1. STABILIZE a highly variable structure (instability is viewed as a marker of incipient change).
  2. EXTEND production of a phoneme from a context where it is currently used appropriately to another where it is not (i.e., structural extension by introduction of a new phonotactic pattern).
  3. EXPAND the system by introducing a new phoneme not currently used by the child (system expansion via the introduction of new phoneme and therefore new contrasts).

In the book we show how all three types of goal can be targeted at once using a mixed horizontal-cycles goal attack strategy.

We also show how you can use Hodson’s cycles approach with the same example case. Barbara Hodson doesn’t like my example very much because I haven’t set the cycle up quite like she does (I always have a little trouble sorting out exactly how to do it from her books because I think that the actual implementation varies with child and historical time and is tied to academic semesters in a way that clinical treatment programs are not in a hospital setting).

Week 20 – Treatment Planning: Service Delivery Models

During weeks 20 and 21 of the course we cover Chapter 8 of my book (Treatment Planning) so we devote a full 9 hours to this material. Even with that amount of time it is difficult to fit it all in because we want the students to be familiar with the relevant concepts, the research evidence and have a chance to practice application of the material to clinical cases. Week 20 is devoted to the first two sections of the chapter, first how to decide if a child requires treatment at all, and then assigning an appropriate and efficient service delivery model. These sections were taught by our clinical coordinator this year. The material to cover includes three models for deciding if a child should receive service (norm-reference approach, medical approach and ICF Framework), provincial regulations for service eligibility, options for service delivery, and the research evidence on the efficacy of those options relative to patient characteristics. If Sophie were to lecture on all that material there would be no time left for an application lab. Therefore, I think the right choice was to require the students to read much of the material independently and focus on an in-class lab. And Sophie designed a brilliant activity for the students – they really enjoyed it and did a good job. She provided small groups of students with a typical schedule for a school speech-pathologist, assuming some number of schools and a communications disorders assistant, along with a list of hypothetical cases and their characteristics, and asked the students to organize their caseload, using DPD Figure 8-1 as a guide to decision making. Figure 8-1 is a flow chart that leads to five decisions for a given case; higher intensity intervention, initiate predictive assessment protocol, lower intensity intervention, monitor academic progress, or discharge.

Of course the chart cannot be applied properly without understanding the background material provided in the chapter along with case examples. Many current studies indicate that students do not read, or more specifically, they read everything but their textbooks! Our textbook is a particular challenge because it is intimidating simply because of its size although it is not difficult to actually read once broached. You might think that these research findings apply only to undergraduate students but I don’t think that it is true – in fact I think that our very bright graduate students at McGill are particularly savvy about doing only what is necessary to get an A whereas I want them to learn to refocus on excellent patient care and never mind their grades. Good SLPs are expected to pay close attention to the ever changing research literature so we need to prepare our students to read that literature. I know a good many SLPs who do not read it at all however so we need to do a better job at the preprofessional level of preparing them to read complex material and to get used to getting information from printed material even when the questions they have are clinical in nature. I think that a lot of our students think that the research literature contains information that is largely irrelevant to clinical practice and that the answers to questions like “How should I organize my case load?” can only be answered by asking another SLP. This is not a good way to move the profession forward (although we have seen some excellent innovations in the past decade such as response-to-intervention). Francoise, my co-author recalls some depressing but hilarious conversations with her clinical supervisors when she asked them to explain why they had organized their school caseloads the way they did – not all SLPs are able to give a coherent answer to that question! So back to the issue of how to get students to read stuff (if not our book, at least other printed sources such as LSHSS). There is some research evidence to suggest that Just in Time teaching is effective – this is an intriguing practice whereby the students have to answer one or two questions about the readings, using the university’s web teaching resources, within a few hours of class onset each week. The questions have to be carefully designed to be counter-intuitive or not easily answered without a deep understanding of the text reading. They must be quick and easy to answer and grade however. I am going to consider this for next year although I am a little worried because I am not sure if this procedure will promote professionalism, i.e., reading for the benefit of the patients rather than external incentives. Possibly using “clickers” in class to get responses to the questions before the practical application lab begins might be an alternative; I have to think about it.

Weeks 18 and 19 – Interpretation of Assessment Data

During weeks 18 and 19 I cover Chapter 7: Nature of Developmental Phonological Disorder. This is the third chapter in Part III: A Holistic Approach to Diagnosis and Treatment Planning. The idea is that the student should consider the scientific literature on speech disorders as a guide to interpreting their assessment results and as an essential foundation for treatment planning. Therefore I encourage them to have finished at least preliminary analysis of the assessment data obtained from their ‘term case’ so that they can apply this rather dry material directly to a clinical case as we study it. In fact, the quality of the students’ term projects this year was amazing, in part because they were able to use the information in this chapter to identify the fundamental underlying areas of difficulty that explained their case’s speech problems and thus chart a course of action with confidence.

Besides the fact that I have a fair few video demonstrations to present, it takes a full six hours to cover the material in Chapter 7 because I introduce multiple models for thinking about subtypes of developmental phonological disorders including those that focus on linguistic symptoms, those that focus on underlying psycholinguistic processes and Larry Shriberg’s new research framework that ties these together with etiological factors (Dr. Shriberg worked with me to create a modified version of the chart for my book that is simpler but you will find a complete version in the paper that I have linked here). I hope that the chapter will stand the test of time, providing the students with a foundation for making sense of the inevitable changes in frameworks and terminology that will be coming – although such change is a constant in our profession, the basic dimensions that we rely on to make sense of heterogeneity within this population have not changed substantially in decades and probably will continue to be influential for some time to come.

It looks like the future will bring a move away from an insistence on discretely packaged subtypes towards a dimensional view as described here for mental health disorders. Ramus, Marshal, Rosen and Van der Lely (2013) described patterns of comorbidity between Specific Language Impairment and Dyslexia with respect to three dimensions: nonphonological language skills (lexical, syntactic and morphological), phonological skills (phonological awareness, rapid naming and verbal short term memory) and phonological representations (accurate production or discrimination of words or nonwords). Given the large overlap between the populations of children with Developmental Phonological Disorders (i.e, primary speech sound disorders) and SLI we can probably fit these children into that framework without adding new dimensions. With the addition of a dimension for verbal praxis we could add MSD-AOS (motor speech disorder-apraxia of speech) to the group, like this:

Dimensional View of SLI and SSD

Teaching – Week 17- Multilinear Analysis

Week 17 covers the larger part of Chapter 6: Speech Sample Analysis, focusing on a comparison of Barbara Hodson’s HAPP-3 analysis and May Bernhardt’s multilinear analysis. Françoise and I show in our book that the HAPP-3 works fine for some children but the multilinear analysis provides a more adequate analysis for children who have problems at the prosodic levels of the phonological hierarchy. The demonstration of this point in the book is quite long so I won’t try to reproduce it here. I would appreciate feedback from anyone who has read the “Amber sample” demonstration however – did it convince you that a multilinear analysis is worthwhile?

Multilinear analysis has the added benefit of being applicable (with some language specific modifications) to French which is important because some of our students end up assessing a child in French each year. In fact, Francoise and I have found that speech sound errors in French-speaking children with phonological disorders are particularly likely to involve interactions between the prosodic and segmental tiers. Françoise has argued that a multilinear analysis is also especially appropriate for the analysis of Spanish and African American English. So, even though we teach phonological process analysis we require our students to conduct a full and detailed multinear analysis of the speech data that they collect from the child that they assess for their term project. We recognize that most practicing clinicians are not going to want to complete this analysis however because it is a time consuming process. Therefore we also teach the students to describe a child’s speech sample from the multilinear perspective after just a quick scan of the data, using a summary Chart (Rvachew DPD Figure 6-8 Quick Multilinear An) to guide their observations. (I include a copy of this chart here but also on the corrections page because there is a small error on the chart in the book.)

Every year I give the students a citation form and free speech sample recorded from a child with a severe phonological delay and ask them to conduct this quick version of the multilinear analysis. This year I was particularly impressed with their responses so I believe that Francoise’s efforts with Chapter 6 were worthwhile – the students have succeeded in learning the basics of a very difficult analysis. A large proportion of the class went beyond the basic summary and used their knowledge of multilinear phonological structure to identify aspects of the child’s underlying phonological knowledge that were not immediately obvious in his surface forms. For example, the students noticed that productions such as ‘duck’ → [gᴧ] and ‘cup’ → [pᴧ] indicated that the child had some knowledge of the feature specifications of the delinked coda consonants. The students had more difficulty integrating the information from the child’s productions of single words with the child’s productions of connected speech however. The child they were studying consistently omitted stops in the coda position of single words but would sometimes include them in connected speech (e.g., in a phrase such as ‘hop on’). It probably would have been too much information for the students to go much beyond the word level in Chapter 6.  This is unfortunate because a primary advantage of multilinear analysis is the ability to explore these links between prosody and segmental output.

Chapter 6 describes some other analysis procedures useful for younger children (phonotactic assessment, phonological mean length of utterance) and for older children (weighted speech sound accuracy measure) but I don’t teach them to the whole class. If one of these analyses is more appropriate for the student’s term project case than multilinear analysis then I teach the method to that particular student. I think that it is best to focus on making sure they all learn multilinear analysis well than providing a brief overview of too many methods so that ultimately they don’t learn any of them.

Week 16 – Speech Sampling

One of the tasks the students complete during their first assessment is to collect a speech sample from a child and then prepare it for further analysis. Even though my students are graduate students, many of them find this to be a very difficult task. It is possible that if they were told to simply go have a conversation with the child they would manage just fine. But somehow, after a 2 hour lecture on measurement issues such as reliability and validity and further discussion of the research on the relative efficacy of different sampling techniques, the students seem to forget how to talk to children! Every year I ask the students to provide a complete gloss of their conversation with the child and a critique of their own technique and an evaluation of the adequacy of their sample. I learned from reading these transcripts that students need very specific teaching in this procedure. One very specific and easy technique that students can learn to use is to repeat what the child says – this helps with the gloss of the sample and ensures that the student follows the child’s lead in the conversation. A good example of the power of this technique is found in a paper by Peterson in her colleagues on their program to teach low-income parents to foster narrative skills in preschool aged children. This paper contains two example conversations that beautifully demonstrate the power of this simple technique. I reproduce them in my teaching slides and always enjoy presenting these when I am teaching the students to collect their speech sample each year. In my teaching slides Week 16 Eliciting the Speech Sample you will find material on threats to the validity of the speech sample from:

Shriberg, L.D. & Kwiatkowski, J. (1985). Continuous speech sampling for phonologic analyses of speech-delayed children. Journal of Speech and Hearing Disorders, 50, 343-334.

and conversations from two parent-child dyads reproduced from:

Peterson, C., Jesso, B., & McCabe, A. (1999). Encouraging narratives in preschoolers: an intervention study. Journal of Child Language, 26, 49-67.

The first is an example conversation from a parent-child dyad that was in the control group. The second is an example conversation from a parent-child dyad that was in the experimental group, recorded one year after participation in the parent education program:

If the students prepare adequately for the experience and use this simple strategy they should be able to collect at least 100 usable words within about 10 minutes of conversation. Developmental Phonological Disorders: Foundations of Clinical Practice provides detailed instructions for evaluating the adequacy of the sample and preparing it for a multilinear phonological analysis.

Teaching – Weeks 14 & 15 – Assessment

I am delighted to be team teaching my winter term course (Phonological Disorders in Children) with our Clinical Coordinator who is a wonderful clinician with lots of experience in the francophone schools milieu. Sophie taught the first two lectures of the term which cover the introduction to Part II and Chapter 5 of the text, Developmental Phonological Disorders: Foundations of Clinical Practice. The main goal of these two weeks of lecture and laboratory material is to prepare the students for their first experience conducting a complete phonological assessment of a child with an identified speech problem. The most wonderful thing about team teaching with the clinical coordinator is that she has arranged for the assessments to be coordinated with the students’ other practicum experiences. She asked the clinical educators to find really interesting cases for the students to assess. That means this year’s term projects should be a lot of fun (the students will be conducting in-depth phonological analyses, selecting treatment goals and designing a treatment program for their own case over the course of the term). My experience with clinicians in the field is that they can sometimes be a bit cavalier about the oral-peripheral examination but Sophie has taken this part of the assessment very seriously. I usually ask the students to complete an abbreviated version that takes about two minutes but she has taught them the full cranial nerve check and asked them to complete the Thoonen et al. maximum performance tasks as I described them in this paper: Rvachew, Hodge and Ohberg (2005). We provided the students with a scoring sheet for collecting and interpreting the data (MPT Data Form 011313)– these interpretative guidelines from Thoonen et al are appropriate for school aged children. For preschool aged children, Thoonen et al.’s scoring criteria require some modification as we suggested in a follow-up paper here: Rvachew, Ohberg and Savage (2006).

Maximum performance task procedures help the SLP to differentiate dysarthria from dyspraxia when diagnosing motor speech disorders in children. Sophie provided the students with preliminary information about these disorders from Margaret Fish’s terrific book (Here’s How to Treat Apraxia of Speech, Plural Publishing) as an introduction. I will cover these topics in more detail when I teach Chapter 7 of my book later in the term.

Teaching – Weeks 10 to 13 – Preschool and School Age Norms

I stumbled across a Sander norms chart posted on Pinterest: http://pinterest.com/pin/1125968626985687/. Of course the chart doesn’t give any information about what the bars on the chart mean and the caption gives a completely wrong impression. Greg Lof did a whole series of wonderful presentations on how to interpret sound-by-sound norms a few years back: http://poti.wikispaces.com/file/view/confusion+about+speech+sound+norms+and+their+use.pdf because he found that so many SLPs were misreading these charts. Anyway, finding the chart reminded me that I never finished my teaching blog for last term and I am about to start teaching a whole new term so I guess I better get up to date. The last 4 weeks of the term were spent on Chapter 4, covering normative data and theories of phonological development. I included a copy of the Sander norms chart in the book except that I annotated it so that the meaning of the bars is printed right on the chart. That way, if students copy the chart (Figure 4-5 in my book) and put it on their office wall – or pinterest or whatever – it is easier to interpret it correctly. The other thing that I do in the book is include a lot of information about how this normative data is collected and aggregated because there is no way to interpret it correctly without understanding these details. I also include information about other types of normative information, collected from a phonological process or multilinear phonology point of view and I discuss a rather ambiguous case from these multiple perspectives. Too bad my book does not include Ballard et al’s new data on the development of lexical stress contrastivity, just published last month http://jslhr.asha.org/cgi/content/abstract/55/6/1822 . However, I stress the importance of attending to the prosodic levels of the phonological hierarchy when describing children’s phonological development throughout chapters 1 to 4.

The last class was a lot of fun. In this class I have small groups describe samples of speech obtained from my former patients, selected to represent a range of severity levels from normally developing through profound speech delay. The groups described the samples using different approaches and judged severity from these various perspectives (phonetic repertoire, sound-by-sound norms, phonological process norms, Goldman-Fristoe Test of Articulation percentile rank etc.). Then when the class regrouped to compare judgments it was revealed that the severity judgments obtained from different normative perspectives didn’t always agree. Some students really enjoyed the exercise because they discovered how much they had learned over the term. Other students were alarmed to find that the process of deciding whether the child’s speech was delayed or not seemed to be, in their words, “subjective” because they would have to use their clinical judgment to reconcile contradictory findings. What if the child obtains a “passing” score on the Goldman-Fristoe but uses phonological processes characteristic of a much younger child? What if the child scores in the severe range on the Goldman-Fristoe but has no nondevelopmental errors? The class discussion was lively. I think that teaching materials often focus too much attention on ideal or straightforward cases because it makes it easier to create and grade exam material. It is the more ambiguous cases that lead to real learning however. I hope that the exercise drove home the point that it is not appropriate to use only one source of data and highlighted the importance of clinical judgment to SLP practice.

Teaching – Week 9 – First Words

Years ago Susan Rafaat and I did a study at the Alberta Children’s Hospital in which we investigated reliability of

SLP ratings of severity of children’s phonological deficit.  We found that reliability was especially poor for the youngest patients (see Figure). More than 15 years have passed and I hope that SLPs are much more knowledgeable about phonological development in early childhood, thus leading to better reliability in the assessment of young children. There are reasons to remain concerned about clinical knowledge for the assessment and treatment of the younger clients however. First, children on our case loads are increasingly even younger than the 30 month cut-off used in this 1995 study and yet text books are still prone to focus on normative data for children three years of age and older. Second, when I advise SLPs I find that it remains common for clinicians to describe the speech produced by toddlers using metrics that are not appropriate for children at this stage of phonological development. If children are at a stage where phonology is still organized at the whole-word level, it is not appropriate to describe their speech in terms of phoneme level error patterns for the purpose of determining the presence of speech delay or identifying treatment goals. Therefore, in our book we devote 16 pages of Chapter 4 to the topic of early phonological development (not to mention the large section on early babbling and speech development covered from the articulatory point of view in Chapter 3). I have a lot of samples of early speech, recorded from children showing normal and delayed developmental trajectories, that the students can practice describing using the two descriptive metrics outlined in Chapter 4: phonetic repertoire and the phonological mean length of utterance. Normative data is available for both of these methods so it is possible to determine whether a toddler is producing speech that is within normal limits with respect to its phonological characteristics if the SLP knows how to describe the child’s productions using these tools. I also introduce the idea of selection and avoidance in early lexical development and show how knowledge of the child’s phonetic preferences can be used to select and order treatment goals for young children. I attach one of the practice activities (Jamie Vivan activities), derived from samples of actual child speech, that I assign to students for practice in the skills taught in the first part of Chapter 4.

Teaching – Week 8 – Theories of Phonological Development

Finally, in week 8 of my course on phonological development we get to where your average course on phonology in a communication disorders program would typically start: Chapter 4 of our book which is the Chapter on Phonological Development. I decided this year to cover the last half of the book first however which is the section on theories of phonological development as well as environmental and biological influences on phonological development. This section is rather heavy going especially given the particular focus on emergentist approaches rather than the more familiar theories of phonological development. Therefore I gave a rather standard lecture which pains me because I try to avoid the kind of lectures where I simply read the book to my students – usually I run my class like a lab and bring in practical activities that allow the students to apply what is learned from the book to a clinical problem.  I am not sure what kind of activity would work best for this section of the book however. It is more common for textbooks to focus on formal models of underlying phonological knowledge and thus the student will find comparisons of different descriptive metrics (phonological process versus optimality theory for example). I have already covered description in Chapter 4 so I didn’t want to ask the students to do any descriptive exercises from different theoretical perspectives. My goal in the latter half of Chapter 4 is to dispel the notion that these phonological structures come built in to the child – I want the student to have some sense of developmental mechanisms that allow the child to learn phonology from the input. What are the expected clinical outcomes of learning about theories of phonological development? For myself, I want the students to understand that phonology does not unfold according to some internally driven innate time table. Rather it emerges from the child’s unique history of interactions among multiple endogenous and exogenous factors. The clinician needs to identify and manipulate all the factors that promote learning of phonology – perceptual, articulatory, lexical, cognitive, social… perhaps I can find some video of children in the process of learning a new phonological skill and have the class speculate on the variables that contributed to each child’s success or failure in the learning of the target skill. It will be hard to find the right bits of video in my lab and I am not sure this will work but it’s worth a try for next year.

Teaching – Week 7 – Development of Speech Motor Control

While at the ASHA convention this year I had the opportunity to hear from people who are reading our book, some who are even using it in their teaching right now, and the chapter that everyone finds to be most notable is Chapter 3: Development of Speech Motor Control, an important topic that is not covered in depth or at all in other texts for courses on phonological development or phonological disorders. The chapter has three sections. The first part describes early speech development from a motoric perspective, the second section describes theories of speech motor control and the final section briefly discusses various influences on the development of speech motor control (structural, neurophysiological, sensory, social and cognitive). Francoise and I also presented a seminar on the Challenge Point Framework and its application to the treatment of speech sound disorders, including apraxia of speech. This seminar involved a lot of interaction with the audience so we had an opportunity to be impressed by the SLPs’ creativity as clinicians but also to discover common gaps in knowledge about speech motor development. I am hoping that my students will finish the year with these gaps filled in.

To help the students learn and apply the model of speech motor control presented in the book I asked them to discuss the model in relation to their own experience teaching someone else a sensory-motor skill (or, if they had no experience teaching, to their own experience learning such a skill). Their responses to this question on the class discussion board were wonderfully detailed and often entertaining, involving such disparate skills as classical singing and kickboxing.  The students showed admirable beginning knowledge of concepts such as inverse internal model and feedforward command. They had a lot of trouble imagining any learning mechanisms beyond imitation and reinforcement learning however. At ASHA I also talked to SLPs who were unaware that children may not be ready to learn via imitation when they had very limited experience with their own vocal apparatus. In Chapter 3 I talk in depth about the “expansion phase” (vocal play) as an important period in vocal development and in Chapter 10 we review Dethorne et al.’s (2009) excellent suggestions for encouraging practice at this level. In Chapter 10 we also mention an interesting approach to physiotherapy, assessed via randomized control trial by Dusing and Harbourne (2010); working from a dynamic systems perspective, they emphasize the role of “variability in postural control to support movement experiences, exploration, and global development.” We think that variability is also essential in the development of speech motor control and this point is often missed. Speech therapy has a tendency to focus on imitative speech drills as a means to promoting precision in the implementation of motor plans. Some children do not have the foundational experiences that support selection of the correct motor plan in the first place. When the child is unready to benefit from an imitative approach to speech therapy the SLP reaches for the nonspeech oral motor exercises and is relieved when these prove easier than speech. But these exercises are unlikely to be helpful because motor learning is context and task specific. There is no better antidote to the desperate adoption of unhelpful therapy procedures than a solid understanding of the underlying theory and the course of normal development.

Teaching – Week 6 – Speech Perception Development

I am weeks behind on this teaching blog now having become bogged down in grading and administration and preparations for the ASHA conference mid-term. My students were feeling the same way, bogged down with the cumulative demands of the term and fully stressed by exams and assignments when I suddenly assigned a 59 page chapter to be read in a single week. This represented a four-fold increase in pace from the first chapter which we covered over the course of a month. As you can imagine I was facing a bit of hostility from students who were not sure that it was really important to read all those details about speech perception development anyway. The publisher wasn’t so sure that a long chapter on speech perception development was necessary either which is understandable I guess because other texts on phonological disorders have very little information. The problem from my perspective is that the little information that does appear in those books is inevitably wrong. Gordon-Brannan and Weiss (2007) provide a rather extreme example in the third edition of their text, Clinical Management of Articulatory and Phonologic Disorders. Only one chapter in the entire book deals with phonological development, and in that chapter there is only one sentence that even mentions speech perception development:  “Perception and discrimination skills are present during infancy” (p. 56). It is no accident that what little is said later in the book about perceptual approaches to intervention is also cursory and wrong! So Francoise and I thought that it was pretty important to give students, and perhaps more importantly their instructors, a solid grounding in speech perception development as an important foundation for phonological development. In particular, the chapter is long because it was necessary to address a number of myths about speech perception development that are long-standing and persistent in the clinical community:

Myth 1: Speech perception skills are innate. In fact speech perception abilities are learned.

Myth 2: Infants are born with the ability to perceive all the phoneme contrasts used in all the languages spoken in the world. Actually there are many phoneme contrasts that infants do not perceive at birth.

Myth 3: There are no individual differences in perceptual ability at birth and thus no need to assess or treat speech perception abilities in older children. It is increasingly clear that some speech, language and reading impairments are associated with genetically determined deficiencies in speech perception that are present at birth and have persistent developmental effects.

Myth 3: Speech perception is just about discriminating one phoneme from another. We take the view that speech perception should be considered in the context of a broader array of phonological processing skills. These include identification and discrimination of a range of speech units including both prosodic and segmental structures as well as related skills such as phonological awareness.

Myth 4: Perceptual development is all over with in the first year of life. Many studies show that speech perception development is a protracted process that continues through late childhood.

Myth 5: Speech perception involves transforming speech input into a representation based on innate phonological categories at which point the acoustically rich and detailed speech input is discarded. Some of the most exciting new research in infant perception is coming from Bob McMurray’s lab at the University of Iowa. McMurray and Aslin (2005) for example show that infants are sensitive to within category detail and suggest that the learning of phonological categories is dependent upon access to this detail in the acoustic input. This research program also tells us a lot about the important role of rich variable inputs in perceptual learning.

Now I suppose I could have made the point in those 5 short paragraphs and skipped the other 58 pages but I take this notion that learning is enhanced by access to the details seriously.  Oddly, just as the science is showing us that learning is dependent upon this process of organizing details abstracted from environmental inputs, it is becoming even more fashionable to propose that teachers should just skip the details because, hey, students can always google them when they need them! Somehow we are supposed to teach creativity and critical thinking in a content-free vacuum. The idea has intuitive appeal because we have all had the experience of remembering the big idea whilst forgetting the underlying the details but of course this doesn’t mean that we could have learned the big idea without first going through the details themselves. Some people think that they can take a short-cut to learning the important concepts: recently a doctoral student told me that he had never read the “middle” of a research paper until it came time to write his dissertation which explains something about the quality of a lot of the paepers that come across my desk as an editor. Anyway, I think this practice of avoiding the details is how we end up with these scandals where fake papers get published (disregarding outright fraud where the editor or reviewer is the person who wrote the fake article!). The reviewers think they know something because they recognize the buzz words but of course, having limited familiarity with the details of the relevant literature, they are completely ignorant and apparently afraid to admit it. So my point is, if you want to know anything about phonology you have to know where it comes from and that means knowing the details of the research about speech perception development in infancy. We have to figure out how to fit this content into the speech pathology curriculum even if there isn’t time for it in the typical 13 week course devoted to phonological disorders. Somewhere in the 6 years it takes to become a speech-language pathologist we have to make sure that the student at least figures out that the infant is not born with an adult-like speech perception system in place.

Teaching – Interpretations of Concepts in Development

During Week 1 when we covered the introduction I gave the students the opportunity to post short blogs interpreting the concepts in Table I-1 that introduces ideas from dynamic systems theory, specifically multicausality, self-organization, phase transition, nonlinearity, heterochronicity, control parameter, coordinative structure, dynamic stability and representational state. I thought you might enjoy the connections that the students made between the concepts and their own experiences or knowledge in other domains. I know I certainly did – the links to videos are particularly entertaining.

Two students tackled nonlinearity. Rachel  linked the concept to the ‘butterfly effect’ and to second language acquisition:

Nonlinearity is defined in the text as “Linear change at the microscopic scale drives change at the macroscopic scale; gradual change at the microscopic scale can lead to large discontinuities at the macroscopic scale; even small differences at the beginning state can lead to large differences in the outcome”.

This concept reminds me of something I learned about in one of my undergraduate courses— the butterfly effect (relevant in chaos theory). Essentially, this effect indicates that a small initial change in a system can lead to very large (and complex) differences in outcomes. The name “butterfly effect” comes from the idea that a butterfly flapping its wing in California, for example, could cause a hurricane in Taiwan several weeks later. This is perhaps an exaggerated example of the idea, but it demonstrates that even a small change in one variable can greatly affect the results or outcome of an event.

I came across a paper titled “Chaos/Complexity Science and Second Language Acquisition” by Diane Larsen-Freeman, which appeared in Applied Linguistics in 1997 (http://applij.oxfordjournals.org/content/18/2/141.full.pdf+html).

Although this paper is somewhat old, it contains a detailed explanation of how language is a dynamic nonlinear system (and how it relates to chaos theory), and the author specifically addresses the fact that language is acquired in a nonlinear way: “There are many interacting factors at play which determine the trajectory of the developing IL: the source language, the target language, the markedness of the LI, the markedness of the L2, the amount and type of input, the amount and type of interaction, the amount and type of feedback received, whether it is acquired in untutored or tutored contexts, etc. Then, too, there is a multitude of interacting factors that have been proposed to determine the degree to which the SLA process will be successful age, aptitude, sociopsychological factors such as motivation and attitude, personality factors, cognitive style, hemisphericity, learning strategies, sex, birth order, interests, etc. (Larsen-Freeman and Long 1991). Perhaps no one of these by itself is a determining factor, the interaction of them, however, has a very profound effect… Further, learning linguistic items is not a linear process—learners do not master one item and then move on to another…” (Larsen-Freeman, 1997)

Therefore, the example of the butterfly effect demonstrates nonlinearity and helps us understand the concept in terms of language acquisition as well— many factors account for the outcome of language acquisition, and presumably a change in any one of them could also change this outcome.

Roberta linked the concept of nonlinearity to her experience with Applied Behavioral Analysis:

Linear change at the microscopic scale drives change at the macroscopic scale; gradual change at the microscopic scale can lead to large discontinuities at the macroscopic scale; even small differences at the beginning state can lead to large differences in the outcome.

This reminds me greatly of some concepts learned during Applied Behavioral Analysis training last summer, which included general training as well as a special focus on working with children with autism spectrum disorders (ASDs).

More specifically, the training included a variety of examples of how each child is unique in his or her set of difficulties and skills, and how there is a plethora of factors that affect the success of an intervention plan, such as: types of reinforcers, physical environment, distracters, therapists, temperament, rapport, motivational stimuli, and many others.

On one hand, a relatively subtle change in such factors (example: providing an opportunity for stimulation for children who engage in self-stimulating behaviours such as hand-clapping or spinning by incorporating a sensory toy before therapy and in intervals during therapy) could have paramount increases in success rate for said children. On the other hand, after a problem behaviour is close to being extinguished, we are told to expect what is called an extinction burst, (something we are told parents have a hard time to deal with as they are worried it is a worsening of the behaviour), where a problem behaviour suddenly increases, despite successful intervention, before it decreases again on its way to full extinction. This can also be followed by spontaneous recovery, the final stage in extinction, where a problem behaviour actually returns temporarily after having stopped completely. This is a time that requires great patience on the part of parents and therapists.

The above examples illustrate my experience with the concept of non-linearity.

To learn more about autism, go to the following link:


To learn more about the Gold Learning Centre for Research, go to the following link:


Roxana also linked to the concept of the extinction burst but in this case, in relation to the concept of the control parameter (with very funny videos to illustrate the concept further!):

“Control parameter” is defined as “the last of many necessary components to reach some critical level or to be recruited to the task in a given context before the system moves on to a new configuration”.

I think that the “extinction burst” procedure often applied by ABA therapists explains what a control parameter is within the dynamic systems theory. In operant conditioning, the term “extinction burst” refers to a temporary increase of an undesired behavior after the reinforcement for that behavior has been removed. If the trainer is patient and consistent, the behavior will eventually go extinct.

Here’s a video about how we can fail to modify behaviour if the extinction burst is not well applied: http://www.youtube.com/watch?v=ZN5PoW7_kdA

In order to picture the function of the behaviour, an ABA therapist has to identify the ABC’s of the situation. In our case, they are the following:

Antecedent: Orange annoys Apple.

Behaviour: Apple reinforces Orange’s behaviour.

Consequence: Orange gets the desired attention.

The result is that Apple will actually increase Orange’s annoying behaviour, by positively reinforcing it. If Apple would completely ignore Orange’s comments, it will elicit an extinction burst, in which Orange would be even more annoying for a while, trying to get the same reinforcement as before. But if Apple is consistent and successful at ignoring the behaviour, Orange will stop the unwanted behaviour and move on to a new topic. It can be seen in the video that Apple has episodes of trying to ignore Orange, but then gives in and reinforces Orange’s behaviour, making it harder to achieve extinction of this behaviour in the future.

In the following video you can see how even a dog applies the principles of the extinction burst procedure, when a toddler throws a tantrum: http://www.youtube.com/watch?feature=endscreen&v=KpSfThUv_pc&NR=1

And here’s a commercial about a very common situation we often witness in supermarkets: http://www.youtube.com/watch?feature=endscreen&v=zM8ym0FqM_k&NR=1

Emily  found a video to illustrate a self-organizing system:

Self organization is defined in the textbook as a phenomenon that occurs when “the constraints binding the constituent parts spontaneously break, creating instability, and then the constituents reform into a new, stable configuration.”

A real-life example of the process of self-organization occurs quite frequently during power outages. In such a situation the stoplights, which normally dictate the rules of the road are not capable of operating. Rather than mass chaos ensuing, a period of instability is followed by the creation of a new stable system, usually a policeman who is stationed at the intersection and directs traffic. Recent evidence however has shown that this police officer may not even be necessary since often in power outages when no policeman is available, busy intersections begin to function in the same way as a four-way stop. The system self-regulates and we don’t even require a machine or a person to dictate whose turn it is to drive. This is an example of a system which spontaneously broke and shortly thereafter reformed into a new stable configuration.


Rachael tied the notion of a self-organizing system to the concept of a coordinative structure and provided a detail description of a ‘feeding frenzy’, illustrated with a dramatic video:

A Coordinative structure is “A functional grouping of compenents that provides dynamic stability though synergistic coordination of the components within the system.”

Synergy, in addition to being the favourite word of any team-leader at every work place training session, means that when A and B work together, the outcome is greater for both parties involved (as opposed to being better for one or the other). By working together, they can obtain something that neither party could alone.

An excellent example of a coodinative structure are feeding frenzies. As a marine biology affectionado, I have witnessed these many times in the Bay of Fundy. One of the most common whales in the Bay of Fundy are minke whales, whose diets consist largely of krill. Many small fish, such as herring, also eat krill. Porpoises, a type of small dolphin, eat a lot of herring, as do many marine birds (ie, gannets and various species of gulls). Bald eagles, another inhabitant of the bay, will often prey on smaller birds (such as kittiwake gulls).

A typical feeding frenzy in the Bay of Funding would work like this: in locations with large concentrations of krill, you will find minke whales close to the surface of the water, and herring below them in the water column. Porpoises feed in pods- often, a couple of porpoises will dive below the school of herring to scare the fish higher up in the water column, where other porpoises will then swim right through, catching as many fish as they can. Since the porpoises are driving the fish to the surface, it makes them a lot easier for marine birds to catch. Finally, it is common to see a bald eagle hovering over the whole thing, in hopes of catching a floating gull off-guard. As you can see, the fact that all of these species need to eat turns out to be benefial for the other species- without necessarily meaning to, they all help each other out (even the poor krill- if the population was not controlled, there would not be enough micronutrients in the ocean either!).

For a really cool example of a feeding frenzy, check this out: http://www.youtube.com/watch?v=RCVAGJEwCqQ . This one takes place in South Africa, so just imagine that the sharks are minke whales, and that the dolphins are porpoises 🙂

As a side note, the school of fish is a good illustration of a self-organized system, too!

Eugénie suggested that a choir behaves like a coordinative structure and provided a wonderful example:

Thinking about how rhythm perception is a dynamic system, I am wondering if I could take this idea to a higher level… What if multiple human beings, experiencing the rhythmic (and acoustic) dynamic system, interact rhythmically together, forming a choir? Is this also considered as a coordinative structure?

Having sung in choirs, I would like to put forward this idea : a choir is like a dynamic system in which each component (singers) has to follow a guide line (score and conductor). Every singer is different and unique, but in the end, it must sound as a whole. None of the voices should be heard individually (unless it’s asked for). A lot of aspects has to be coordinated : pitch, timbre, rhythm, dynamics, style, etc. Also, if one of the component (singer) makes a mistake, it has to come back together. The fact that, usually, singers around you sing the same part as you must obviously help you get back on track and so, not stop the system.

In short, I think a choir (or orchestra) could be interpreted as a coordinative dynamic system, especially when is it constituted of professionals!

To show you what I mean, here is a video of the SMAM choir (which includes some of my good friends!). It is an early music peace for 6 voices. Notice the interactions between the different parts and even the ‘dynamic system’ position they stand in. In early music, more than in the classical or romantic era, it was important to try to sound as unified as possible.


Raphaelle interpreted the concept of dynamic stability by linking to her own experience in addition to a fun video:

“Dynamic stability” occurs when the coordinative structure flexibly self-adjusts to the ongoing demands of the task context.

This concept evokes memories of my children learning how to ride a bicycle. And… memories of me running behind them until they learnt 😉 The learning was progressive and different for each child. Each had to learn to feel what she/he needed to do to keep her/his balance while moving forward. Once she/he acquired the knowledge, there was no going back.

The bicycle is unstable when upright and motionless but it can relatively easily (once learnt) be balanced when in movement. When dynamic -like in movement- is it more stable. Dynamic stability.

High on the expertise scale, there is this African man who dances with his bicycle (http://www.youtube.com/watch?v=9zThF47FOXQ). The bicycle adjusts to the ongoing demands of the dancer’s movements. The dancer’s keeps creating instability and adjusting the position of the bicycle to regain his balance.

Claudette  provided an interesting example of heterochronicity:

According to Table I-1, heterochronicity is described as

“The constituent components in the system change according to their own separate but interacting timetables.”

One example of this is chronobiology, which is the study of biological rhythms that enable us to function mentally and physically. The following link indicates that circadian rhythms and homeostatic sleep-wake cycles interact to modulate human cognitive performance. Although these cycles operate on their own separate rhythm, they also interact together in such a way so as to optimize our cognitive processes.


Finally Vanessa wrote some paragraphs about multicausality:

Multicausality is defined as “New behaviours arise from the complex interactions among multiple endogenous and exogenous components that have causal equivalence.” As I understand it, multicausality underlies overall language development and brings some mediation to the nature/nurture debate.

Human language consists of several characteristics that differentiate it from other forms of communication. In this sense, language must be species-specific, an ability that is innate to the human being. Two innate factors that contribute to language development have been underlined: the so-called language acquisition device (LAD) and universal grammar (UG). The LAD is thought to be a domain-specific mechanism in our language faculty that allows for the acquisition of language in humans. UG is the idea of an innate, abstract set of linguistic principles and parameters that dictate what is possible in human language, such that all languages share some common underlying features.

However, these endogenous mechanisms (nature) are not enough to ensure the proper development of language in children. Exogenous factors in the environment (nurture) play an equally important role. Studying children who are deprived of sufficient linguistic input during the critical period of language development demonstrates that without adequate exposure to the language being learned, language acquisition does not take place properly. One popular example is the case of Genie, who was deprived of linguistic input until the age of 13. She acquired words and was able to communicate verbally but did not fully develop her morphological or syntactic knowledge. Therefore, it can be said that when it comes to language development, endogenous and exogenous factors are equally important for ensure successful acquisition.

Teaching – Weeks 4 and 5 – Multilinear Phonology

My students are writing their midterm exam today which is Week 6 of this course (if you discount the missed lecture due to Canadian Thanksgiving). While I proctor I have a little time to catch up on my teaching blog. During Weeks 4 and 5 we covered Multilinear Phonology (and a very brief run through phonological processes so that they will be able to talk to their clinical supervisors during their first practicum!). Bernhardt and Stoel-Gammon (1994) provided an excellent introduction to multilinear (i.e., nonlinear) phonology for a clinical audience but it is structured so as to emphasize points of intersection between the phonological process and nonlinear approach. I believe that the paper takes a case study that served to demonstrate the phonological process approach in Stoel-Gammon’s text (Normal and disordered phonology in children) and then they re-analyzed the child’s phonology from the multilinear perspective. This is a sensible way of introducing a something very new since it creates a bridge between the current state of the profession’s knowledge and the new approach. However, the disadvantage is that it is not completely clear why you would want to go to the trouble of crossing the bridge! For example, Table 5 in their paper indicates that the phonological process approach leads to the selection of Velar Fronting as a treatment target for the example case whereas the multilinear analysis results in the selection of Dorsal as a target – in both cases, one would work on words that contain /k,g/ in the syllable final position. This is an oversimplification because the multilinear analysis highlighted the prosodic difficulties in the child’s speech and moved the targeting of Dorsal until later in the treatment program, after the more elementary prosodic concerns were dealt with. None-the-less, I think that SLPs may often see these newfangled linguistic analyses as just a new set of labels for familiar concepts.   The challenge is to show that the analyses reveal aspects of the child’s phonology that are hidden with the phonological process approach. For me there are two really important innovations with multilinear phonology (relative to the phonological process approach) that I want to get across to students.

The first advantage of multilinear phonology is that there is no assumption that the underlying representation is adult-like. Wrapped up in this important point is the second advantage and that is the hierarchical structure of the representation and the way that default structures, delinking and spreading rules interact with this hierarchical structure to explain the child’s output. However, the idea that the child’s representation is unique and not a copy of the adult version is fundamental to me. The SLP has to figure out what the representation is as well as describe the rules that are operating on the representation to yield the output. Sometimes this results in a situation that is exactly parallel to a phonological process but not always. Returning to the common process of velar fronting for example there are cases where the child has knowledge of the Coronal-Dorsal contrast but the child delinks the Dorsal node in some contexts, in which case we can say that the child is “fronting velars” – it doesn’t make any sense to say that the child is “fronting” velars unless you can prove that the child has knowledge of the contrast in the first place. Here is a hypothetical example of a common pattern where this might be the case:

(1)     ‘key’ [ti]             ‘cup’ [tʌp]          ‘cough’ [tɔf]

‘peak’ [pik]        ‘puck’ [pʌk]       ‘knock’ [nɑk]

‘tea’ [ti]               ‘ton’ [tʌn]          ‘tall’ [tɑl]

‘Pete’ [pit]          ‘putt’ [pʌt]         ‘pot’ [pɑt]

In example (1), especially if the child has adequate perceptual knowledge of the /t/-/k/ contrast, ‘fronting velars’ and ‘delinking of the Dorsal node’ are equivalent descriptions of the mismatch that we observe between the adult target and the child’s production in the onset. However, many other possibilities are observed in child speech that might be described as ‘velar fronting’ but the underlying representation is not adult-like at all. In the following four hypothetical cases (especially if the child demonstrates poor perceptual knowledge of the contrast), it would be better to hypothesize that the segment in onset and coda is underspecified for place in the underlying representation regardless of the adult target:

(2)    ‘key’ [ti]              ‘cup’ [tʌp]          ‘cough’ [tɔf]

‘peak’ [pit]         ‘puck’ [pʌt]        ‘knock’ [nɑt]

‘tea’ [ti]               ‘ton’ [tʌn]          ‘tall’ [tɑl]

‘Pete’ [pit]         ‘putt’ [pʌt]         ‘pot’ [pɑt]

(3)    ‘key’ [ti],[ki]      ‘cup’ [tʌp], [kʌp]            ‘cough’ [tɔf], [kɔf]

‘peak’ [pit],[pik]  ‘puck’ [pʌt], [pʌk]    ‘knock’ [nɑt],[nɑk]

‘tea’ [ti], [ki]     ‘ton’ [tʌn], [kʌn]            ‘tall’ [tɑl], [kɑl]

‘Pete’ [pit], [pik]‘putt’ [pʌt, [pʌk]        ‘pot’ [pɑt], [pɑk]

(4)    ‘key’ [ti]              ‘cup’ [tʌp]          ‘cough’ [tɔf]

‘peak’ [pik]        ‘puck’ [pʌk]       ‘knock’ [nɑk]

‘tea’ [ti]               ‘ton’ [tʌn]          ‘tall’ [tɑl]

‘Pete’ [pik]        ‘putt’ [pʌk]        ‘pot’ [pɑk]

(5)    ‘key’ [ti]              ‘cup’ [tʌp]          ‘cough’ [kɔf]

‘peak’ [pit]         ‘puck’ [pʌt]        ‘knock’ [nɑk]

‘tea’ [ti]               ‘ton’ [tʌn]          ‘tall’ [kɑl]

‘Pete’ [pik]        ‘putt’ [pʌt]         ‘pot’ [pɑk]

Asking the student to map out the underlying representations and then consider how the child gets from the underlying structure to the observed output (as they are doing on their exams today) forces a deeper level of thought about the child’s system than just counting up the frequency of ‘fronting’ or ‘backing’ that is observed at the surface level. Reasonable hypotheses for these cases are as follows: In example two I would assume that the child has no phonological or articulatory knowledge of /k/ or the Coronal-Dorsal contrast in the consonant system and the onsets and codas in all words are unspecified for place in the underlying representation, defaulting to Coronal in all cases. Example 3 is more difficult and I would advise instrumental investigations of the child’s speech in this case to determine if the child is actually producing the phones [t,k] inconsistently for the two targets or whether the listener hears the [t,k] variably in response to an ambiguous and probably atypical articulatory gesture on the part of the child. In any case this child does not have clear phonological knowledge of the /t/-/k/ contrast. Examples (4) and (5) demonstrate allophonic rules as discussed for Week 3. In example 3, [t] is an allophone of /t/ that appears only in syllable onsets whereas [k] is an allophone of /t/ that appears in only in syllable codas. Example 5 again involves a target in the underlying representation that is unspecified for place and Dorsal is obtained via spreading from the vowel in the [ɑ,ʌ,ɔ] contexts. Identifying the pattern in Examples 4 and 5 is facilitated by thinking about the hierarchical structure of phonological representations.

In our book, Developmental Phonological Disorders: Foundations of Clinical Practice, we promote an approach to speech therapy that involves increasing the child’s knowledge in the perceptual, phonological and articulatory domains, bringing the child’s knowledge at all three levels of representation gradually into alignment. The students’ ability to identify complex patterns in child speech varies greatly within the class at this early point in the course. Their sensitivity to the possibility of unique underlying representations and the need to consider knowledge in all three domains is fully established however so I am well pleased with their progress in the class so far. The other way that I am pleased with the class is that they have discovered several mistakes in this part of the book. Some of the errors are very minor but one is BIG FREAKING MISTAKE so I have made available a corrected figure for download in a new page called DPD Corrections.

Another excellent tutorial on multilinear phonology:

Bernhardt, B., Stemberger, J. P., & Major, E. M. (2006). General and nonlinear phonological intervention perspectives for a child with resistent phonological impairment. Advances in Speech-Language Pathology, 8, 190-206.

Teaching – Week 3 –Segmental Phonological Knowledge

The second topic that I cover in Week 3 is one of the most difficult to be covered all term. The students think that Segmental Phonological Knowledge will be relatively easy because initially segments seem rather familiar to them and they all know the definition of a phoneme from their freshman year. However the analyses borrowed from generative phonology – identifying allophonic and neutralization rules and phonotactic and inventory constraints – are difficult to do. I introduce this very early in the course because the subtext is the foundation of everything that comes after – the child’s phonological system is not just a simplified version of the adult system. We must approach the description of the child’s phonology as if the child had a unique linguistic system. One reason that I don’t like the phonological process approach (and more-or-less refuse to teach it) is that the underlying assumption is that the child has adult-like underlying representations. In many cases this is not true, especially when assessing children with developmental phonological disorders. Because the students have difficulty with these analyses I provide several examples in the text, a home practice activity that is very difficult and an easier in class activity to work through as a group before providing additional practice on the midterm exam. I try to make the practice activities fairly realistic – in other words, the data is not made of up minimal pairs. Rather the samples represent words a young child might actually say.The challenge for the SLP is to hypothesize the “rules” that might account for the speech patterns observed in the sample and then construct lists of minimal pair words to test those hypotheses. I have attached some of these exercises for anyone to use with their classes. I have not attached my answers because I don’t want them to be openly available before assigning these activities again next year!

Bobby Sample Practice Activity for Segmental Analyses

In Class Practice Activity for Segmental Phonological Knowledge

Teaching – Week 3 – Describing Perceptual Knowledge

In Week 3 I covered two topics from Chapter 1 of Developmental Phonological Disorders: Foundations of Clinical Practice. First I briefly covered units of analysis when describing the child’s perceptual knowledge of the phonological system being. This includes basic knowledge of acoustic cues and a little bit about categorical perception.  I would prefer to not teach phonological development until after the students have taken their speech science course but this is not possible so I am forced to introduce a little bit of basic science very early in the course. Just enough to interpret a spectrogram and relate acoustic features to phonetic units. I was slow to realize that the many students could not find the colour Plates 2 and 3 that are inserted right into the middle of the book and therefore they had no idea what they were reading about! I’ll know next year to show them where the colour plates are on the first day of class. The most important concept that I want the students to understand is that phonological categories are imposed on a parametric acoustic space and that each individual listener has their own strategy for doing this. I had all the students complete the “Sue” module of the SAILS program this year. This turned out to be complicated because even the “new” version of the software will not run on the new 64 bit computers so about half the class piled into my lab to do it on the old computers there. The conditions were not ideal, a lot of background noise and different sets of headphones being passed around. I was worried that there would be a lot of random error in the data set. But in fact there was very little random error-rather the error was quite systematic. Quite a few students identified the three tokens transcribed as /θu/ as being correct exemplars of “Sue” and identified one of the correct “Sue” tokens as being incorrect. Looking at the acoustic characteristics of these stimuli it is very clear that these listeners weight duration of the fricative as the important cue to the /s/-/θ/ distinction. Because SAILS mostly contrasts correct adult productions of “Sue” with incorrect child productions, the temporal and spectral cues are a mishmash.  The children who substitute [θ] for /s/ actually produce the segment with longer duration than the adult correct productions of /s/. The SLP students who weight the duration cue highly wrongly identified the child [θ] productions as /s/ even though they were not produced with the [+groove] feature. The correct “Sue” token that was frequently identified as being incorrect has a centroid of 5985 Hz, within the /s/ range, but it is quite short at 195ms. Once again, the students’ responses indicate a heavy reliance on the duration cue. Heavy reliance on the duration cue is not a problem when listening to normal adult speech when the duration and spectral cues to these phoneme categories are correlated in a predictable fashion. The students need more experience listening to child speech and attending specifically to the spectral characteristics of the fricative if they are to be good at assessment and treatment of phonological disorders.

As an aside, SAILS can be obtained for free from http://mcgill.flintbox.com/public/project/6237/ .  The license is actually free. Be sure to download the ‘backgrounder’. As I said however, it won’t run on 64 bit computers. I am slowly working towards a new tablet version but it will be awhile before that is ready for distribution.

Teaching – Week 2–Describing Articulatory Knowledge

After introducing the students to the concept of multiple levels of phonological representation, the first week of class is devoted to methods of describing articulatory knowledge of the sound system. Each week I want to highlight a few particularly important concepts from the DPD text and then provide practice in one clinically relevant skill. Since the lab will cover phonetic transcription over the course of the term (and the students typically had good exposure to the IPA system in their undergraduate programs) it is not necessary for me to spend a lot of time reviewing the IPA chart and the place-manner-voice characteristics associated with each segment and symbol (students can do this on their own time from the course resources in any case). The primary concept to stress at this point is “motor equivalence” – the idea that different combinations of articulatory gestures can have the same acoustic outcomes. I also like the students to understand that phonetic transcription is a process of imposing discrete categories on an essentially continuous underlying phonetic substrate. One problem with the teaching environment at McGill is that we don’t have a phonetics lab (or any kind of lab actually) so it is not easy for students to experiment with visualizing their own articulatory gestures or to explore the acoustic outcomes of their own speech movements. I have recently discovered however that you can buy an ultrasound for only $5000 which is pretty amazing. You get some very easy to install software and a probe that plugs into the USB port of an ordinary laptop. This is definitely a ‘you get what you pay for’ kind of situation because the images are not anywhere near as clear as those provided by a more expensive system (see Penelope Bascalvi’s terrific examples created in Gick’s lab on the UBC website here. However, the less expensive unit does provide a view of tongue gestures and Jon Preston has been evaluating this system in single subject trials for the treatment of /ɹ/ misarticulation in children.

In addition to understanding the somewhat arbritrary classification of place-manner-voice characteristics of phones in the IPA system relative to the varied articulatory gestures that may occur in running speech, I also like the students to practice transcribing toddler speech and deriving a phonetic repertoire in the first week of class. It’s completely amazing how many parents post audiovisual recordings of their toddlers on youtube so there is a wealth of material available to practice this crucial clinical skill. This way the students learn about the limitations and the benefits of phonetic transcription as a clinical tool in the first week of class.

Teaching – Week 1-Introduction

In Week 1 the assigned reading is the Introduction to Part I of Developmental Phonological Disorders, a 9 page section at the beginning of the book called Phonology from a Developmental Perspective.  My students are graduate students and I am hoping that they will notice that this book is different from many of their undergraduate text books. I point out one important difference by stating clearly that “throughout, we have a particular point of view and have not attempted to present a variety of theoretical perspectives as in a survey text” (p.6). In the Introduction to Part I we outline this point of view and introduce some concepts that will re-occur throughout the text. The point however is not to insist that every reader adopt our point of view. The more important point is that competent and professional practice as a speech-language pathologist requires that the practitioner have a point of view! In my experience, this is a novel idea for many young students who have been taught that it is immoral to be “judgmental” and disrespectful to conclude that any person or position might be wrong. And yet it is the role of a professional to make judgments (does this person have delayed speech warranting a government funded service?) and then decide on the correct course of action (a proscribed period and type of speech therapy) while warning against charlatans (no, a dose of oxytocin is not likely to cure the problem!).

The student must come to some understanding of the theoretical foundation that will underpin their own practice. This understanding will not and should not come all at once since becoming a professional is a developmental process that takes some time and involves a myriad of influences. The student may, over time, come to adopt a position very close to that espoused in our book. The student may develop a theoretical perspective that melds ours with other perspectives encountered elsewhere (given the certainty of future gains in scientific knowledge of phonological development, we would hope that this would be the case). The student may also settle into a point of view that is diametrically opposed to that put forward in our book. It doesn’t matter so long as the student brings to conscious awareness their theoretical viewpoints and the links between their underlying philosophy and their professional decisions.

Even though many of the students may not have thought carefully about what their own theory of development is, we can be sure that they have one because everyone does. What are the characteristics of that theory likely to be? Smith and Thelen (1993) state that a theory of development “must account for the induction of novel structures and behavioural forms; It must account for similarities in development among members of a species as well as individual differences between members of the species.” As undergraduates it is very likely that they encountered Chomsky’s theory of language development and thus they will be comfortable with single cause and nativist theories of development (Dorothy Bishop has written an excellent summary and critique of this perspective on her blog; http://deevybee.blogspot.ca/  Monday 3 September 2012 What Chomsky doesn’t get about child language). In any case the students will have absorbed the prevailing zeitgeist in which a rather unsophisticated form of biological determinism plays a dominant role. In our book we take a very different view, specifically that “speech development is viewed as the emergent product of cross-domain knowledge, environmental inputs, and the social and functional needs of the child” (p. 3). I want the students to begin thinking about complexity, nonlinearity, and multicausality in development on the first day. How you do that with your students may vary.

When I was teaching undergraduates I would ask the students to discuss with their neighbor similarities and differences between themselves and a sibling (or parent or other relative). This activity would be followed by a class discussion about all the possible factors that might account for those similarities and differences. When teaching doctoral candidates in a graduate student seminar, an exploration of the many studies conducted by Smith and colleagues related to the A-not-B task described in the book would be a good introduction to many important concepts in development that reoccur in the text. When teaching graduate SLP students I like to introduce them directly to the clinical context of the course and therefore I begin by introducing four samples of speech: a pre- and post-treatment sample from a child who made good progress in therapy and a pre- and post-treatment sample of speech from a child who made poor progress in therapy. These samples provide a rich foundation for thinking about causes of similarities and differences in development as well as the “induction of novel forms”. Why have the children not achieved intelligible speech like other four-year-olds? What are all of the possible reasons that one child made good progress in therapy and the other did not? What might be done to help the second child acquire the speech sounds that will lead to intelligible speech? When facilitating this discussion I will have the opportunity to point out when the discussion relates to the concepts presented in Table I-1. I also have an opportunity to observe how long it takes the class to suggest that (with or without prompting) that the speech-language pathologist and his or her treatment decisions may have played a role in the different treatment outcomes experienced by the two children!

Further to the concepts presented in Table I-1, one thing I am not going to do is show power point slides representing the definitions found in the table and then turn them into multiple choice exam questions at some later date. It is my intention that the students gradually develop a sense of these difficult concepts as they re-experience them repeatedly in new contexts over the course of the year. I have started an on-line discussion group for my class so that we can all share information and experiences that exemplify these concepts. I enjoy finding examples of the functioning of dynamic systems in operation. For example, a video of a murmeration that went viral on the internet yesterday was an excellent (and stunning) example of emergent structure: http://vimeo.com/31158841. I will note other examples if they arise over the term and invite readers to make suggestions as well.

Orientation Week – Course Goals

I am well pleased to be heading into orientation week here at McGill with my syllabus ready and all my course materials prepared in our on-line teaching system. This is my favourite time of year when I get to meet all the new students as well as find out how the returning students made out with their summer practicums. This year I am welcoming one new doctoral student and a new post-doctoral fellow into my lab. I will also greet 28 first year graduate students enrolled in our applied master’s degree program who will take my phonological development class in fall term and phonological disorders class in winter term. One of the great things about teaching at McGill (besides the small class size) is the fact that there are two classes on phonology and I don’t have to squish all the material into one 13 week class. We are the only program in Canada to provide a 3 credit course on phonological development prior to the course in phonological disorders. It’s not clear to me why it is universally recognized that language disorders cannot be taught until after the students have learned about language development but the teaching of phonology is so frequently shorted in the speech-language pathology curriculum. If any readers have ideas about that, I’d love to read your comments! Ray Kent, in the forward to our book, explained eloquently why it is necessary to do a thorough job of teaching phonological disorders in the speech-language pathology curriculum:

Developmental phonological disorders (DPD) are at the center of the profession of speech-language pathology. These disorders have high prevalence in children and frequently accompany other kinds of communicative disorders. A course on DPD is among the first clinical courses in most academic curricula. A sure knowledge of DPD is leverage for understanding other disorders of communication and also for understanding the clinical process itself. The early appearance of the DPD course in a curriculum does not mean it is an easy subject matter. On the contrary, DPD necessarily demands background information from several fields of study, because phonological development, typical or atypical, is a process that draws on a range of knowledge and skills in the developing child. DPD unfolds across several domains of study, including speech perception, speech motor control, phonology and other aspects of linguistics, acoustics, and child development, to name a few. Children are complex, and so is phonological development.

In the fall term I will take the students through the first four chapters of ‘Developmental Phonological Disorders: Foundations of Clinical Practice’. One of my doctoral students, teaching the laboratory component of the course, will teach selected chapters of Shriberg and Kent’s ‘Clinical Phonetics’. The goals of the phonological development class will be as follows:

The purpose of this course is to provide students with a firm foundation for clinical decision making when assessing and treating developmental phonological disorders. It is expected that you will learn to:

  • reliably transcribe samples of speech recorded from children with delayed phonological skills;
  • understand the multiple representations approach to phonological development and the implications of this approach for the description of phonological knowledge;
  • describe the phonetic and phonological characteristics of speech samples at multiple levels of the phonological      hierarchy;
  • describe the normal course of phonetic/phonological development from infancy through late childhood.
  • Identify the primary variables that impact the course of phonological development from infancy through late childhood.
  • determine whether a child’s level of speech accuracy is within normal limits.

Every year is a new adventure and I look forward to blogging about this years’ experience in this space.

Leave a comment

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: