Share this article on:

“Language Construction in an Autistic Child: Thoughts Regarding Language Acquisition and Language Therapy” Translation, Update, and Commentary on a 1977 Case Report

Thomke, Hellmut PhD*; Boser, Katharina PhD

Cognitive and Behavioral Neurology: September 2011 - Volume 24 - Issue 3 - p 156–167
doi: 10.1097/WNN.0b013e3182351276
Historical Update

A 1977 Swiss case study is presented in English translation: a mute child with infantile autism is taught to speak starting at the relatively late age of 6. The author, who is the primary therapist and the child's father, details the conditioning procedure, discusses theoretical considerations in speech acquisition, and outlines the limits of the training. The author and translator update the child's status and add commentary.

*Professor Emeritus of German Literature and Language, University of Bern, Bern, Switzerland

Individual Differences in Learning and Boser Educational Technology, Ellicott City, MD

HT is the author of the original paper, updates, and some commentary. KB is the translator of the original paper and updates, and author of some commentary.

This research was supported by the Cognitive Neurology Gift Fund and by the Therapeutic Cognitive Neuroscience endowment and gift funds.

The authors declare no conflicts of interest.

Reprints: Katharina Boser, PhD, Individual Differences in Learning and Boser Educational Technology, 4120 Sears House Ct, Ellicott City, MD 21043 (e-mail:

Received August 18, 2011

Accepted August 18, 2011

Back to Top | Article Outline


Katharina Boser (KB)

The core of this paper is KB's translation of a 1977 article by Thomke,1 then a professor of German Literature and Language at the University of Bern, Switzerland. Dr Thomke was reporting on his use of operant conditioning to teach language to his son, Heiner, who had nonverbal autism. Both father and son are identified by name here because they were named in the original publication.

At the time that Heiner was diagnosed, many contemporary accounts of improved speech production in autism were claiming that if affected children could not speak by age 5, they would never be able to learn to speak.2–4 The perception of a “critical period” for language development has continued despite evidence of older children acquiring speech. In fact, a 2009 literature review by Pickett et al5 documents 14 studies in which 15 children aged 5 or older, including Heiner Thomke, were trained to speak spontaneous multiword or phrasal utterances.3,6–17 The reviewed studies used many methods to aid speech/language development, among them sign language and special prompting techniques. Higher pretraining intelligence quotient scores, ability to imitate, and other individual characteristics may have given some of the children a potential advantage over Heiner. Pickett et al5 confirmed that longer therapy generally correlates with a higher level of language attainment, and specifically that children need at least 3 years of therapy to reach the highest level, which is spontaneous phrases. (Heiner had 2 years and 3 months of intense therapy.) The study also showed that longer treatment did not always correlate with better outcome. What stands out in the article is that of all the children studied, only 9 whose diagnosis and treatment were well documented were aged 6 or older when instruction began and learned to produce phrases. Heiner Thomke was one of those 9.

Dr Thomke writes in remarkable detail about his son's therapy, progression, and results. The author's background as a language philosopher has given him special insight into not only his child's language abilities but his own theoretical arguments for the techniques that he used to train Heiner. Because a father is writing about his own son, the reflections must be somewhat subjective. Further, the lack of a research design precludes scientific proof that the behavioral conditioning was responsible for the improvements in Heiner's speech. But the paper gives us a snapshot of therapies for children with autism in Switzerland in the late 1960s and 1970s.

The therapies outlined are not intended to be prescriptive or even replicable, but rather descriptive. Several resemble techniques used today. For example, Dr Thomke's redirecting therapy in response to Heiner's interests is a part of Pivotal Response Therapy.9 The father's use of pictures and objects to help Heiner learn to communicate presages the Picture Exchange Communication System.18 And his technique of training and reinforcing one small skill until it was firmly learned, and then building on it in the next skill taught, is known as “shaping and chaining.”19

Dr Thomke opens his paper by describing the confusion and hope that he and his wife felt when they were confronted with the need to teach their 6-year-old mute son to speak. They are frustrated not just by the difficulty of the task but by the lack of human and written resources available to guide them. Dr Thomke develops his operant training system through trial and error, studying the psycholinguistic and developmental psychology literature, and recognizing the reasoning behind things that he does instinctively. He chastises himself for not starting Heiner's therapy earlier and not being able to give as much time to it as he would wish. Not surprisingly, he also has mixed feelings about the success of his effort. But through his work, Heiner gained functional language and kept learning after his formal training ended. Although Heiner's expressive and receptive language skills have remained limited and he never learned to read or write, his ability to speak has enabled him to be more independent than he could have been otherwise, and his receptive language developed to the point that he has long been able to enjoy listening to recorded stories. Even this partial success has greatly improved his quality of life.

Thanks to Dr Thomke's generous willingness to share and revisit his paper, we have been able to augment Heiner's early history, therapy, and results, and add updates about him and his abilities over the 37 years since his intensive language training ended at age 8. [New material added within the translated paper is shown in brackets.] Dr Thomke also approved the addition of several section headers to mark natural transitions in the text, and the deletion of 3 footnotes not considered crucial.

In 2011, we still seek best practices to help autistic children learn language. In the United States, some school districts have failed to implement evidence-based treatments for students with autism spectrum disorder,20–23 many because of a lack of training.24 The use of applied behavioral analysis for autism in Switzerland has grown only since about 2001, largely because 2 mothers advocated for it. Dr Thomke's 1977 psycholinguistic approach serves the whole child, using sound, form, and meaning, based on a profound knowledge of language structure and philosophy. If his intense training succeeded in enabling a mute older child to speak in sentences, we may be able to apply his methods to younger children and to children with milder autism spectrum disorders.

Back to Top | Article Outline

Language Construction in an Autistic Child: Thoughts Regarding Language Acquisition and Language Therapy 1

(Originally published as: Sprachaufbau bei einem autistischen Kind: Überlegungen zum Spracherwerb und zur Sprachtherapie. Schweizerische Zeitschrift für Psychologie und ihre Anwendungen. 1977;36:1–18.1)

Back to Top | Article Outline

Hellmut Thomke

Back to Top | Article Outline

Translation from the original German by KB

Language acquisition is in many ways still such a great puzzle that the person who stands before the task of teaching language to a child who has not learned to speak on its own, hardly knows how he might bring this about and where he might begin. Professionals and academic texts on language therapy also leave him for the most part “in the lurch.”

Similarly, parents of a nonverbal autistic child are left with feelings of confusion combined with an intuitive sense that if they dare an attempt, perhaps a way will open itself to them; however, without strong methodological experience or scientifically supported ideas about the best approach, this is not something achievable for the long run, something my experience with Heiner, my own child, taught me.

How intuition and reflection can play a role, and how I came to draw conclusions about language acquisition by examining a single case of pathology—my own son—may be illustrated by the following example: when in Heiner's presence, our family decided quite spontaneously to more clearly stress (intone) the most meaningful words, increasing the importance of sentence melody. It was only after quite a period of time that we began to realize that Heiner might perceive our speech more easily through this change of sentence stress. This thought was based on ideas in the newer medical literature about early autism, in which such children were found also to have a higher probability of perceptual impairments, which make the normal development of symbolic abilities (including language) impossible. With this notion arose the idea that perceptual abilities might actually be a prerequisite for language acquisition. Later, I became convinced of the importance of auditory stress and sentence melody by studying the psycholinguistic literature, in which the stress of the tonal gestalt of sentences and the intonation of utterances are hallmarks of the preverbal phase of language acquisition (references 25 and 26, among others). In addition, I determined that by overemphasizing intonation, we created in Heiner a preference for so-called “content” words over “function” words. [“Content” words are words with a real-world referent: nouns, verbs, adjectives, adverbs. “Function” words have no real-world referent; they have meaning only within syntax, eg, “if,” “and,” “but,” and “that.”] “Content” words not only stood out most prominently throughout Heiner's language instruction, but they also take priority in the early language acquisition phases of nonlanguage-impaired children. Thus, it seemed an important thought that the sound patterns of language provided particular direction for Heiner as he learned specific types of speech acts.

[Heiner is the youngest of my 3 children. His sister, Elke, was born in 1959. She was healthy until age 22 months, when she suffered a serious fall, with a great loss of blood. This led to excessive growth of her tongue, which made it difficult for her to speak. At age 4, probably also related to the fall, she developed temporal lobe epilepsy. A seizure at age 15 caused her to drown. My second child, Roland, born in 1962, grew up normal and healthy.]

[Born in 1966,] Heiner appeared in his first year to develop normally, both physically and cognitively. Although his babbling phase was not very productive, at 10 months he was able to produce the sound sequences “mama,” “papa,” “tatata,” and “nan.” In his second year, however, his development halted [and, by 18 months, he could not produce a single sound]. At the time, there did not seem to be a specific reason for Heiner's muteness. Soon early signs of autism appeared. He hardly reacted to speech, he avoided eye contact, and he did not want to be touched. He preferred to watch things that spun and twirled, like wheels. Only somewhat later did he show a fear of change. Strangers thought he might be deaf, although his hearing was tested many times and was always normal. He often reacted to very faint sounds, but generally tried to escape from loud noises. Yet, from an early age, he was engaged by music.

He received the diagnosis of “autistic tendencies, mute” at about 2 years of age. Not long thereafter, the diagnosis was revised to “childhood autism.” Later he was described as a “schwer” (severe) case. [The initial diagnosis was made in 1968 by Dr Hanspeter Matthys, then the only privately practicing child psychiatrist in the State of Bern. In 1970, the diagnosis was confirmed at the pediatric clinic at the University Hospital of Bern.] It was a typical case of Kanner's syndrome. Heiner also had digestive difficulties resulting from unexplained food allergies and perhaps a difficulty in tolerating sucrose. (This made it extraordinarily difficult to find an appropriately motivating reinforcer during the conditioning training.) Multiple attempts to modify his behavior through medication failed. However, thioridazine (Mellaril) and pyritinol (Encephabol) seemed to bring about some improvement at first.

[More on medicines and medical care: Starting shortly before Heiner turned 2 years old, and continuing until about his third birthday, we tried many medicines to treat his autistic behaviors and severe insomnia. There were so many medicines that I can no longer name them all, but I do remember these: the antipsychotic drug thioridazine (Mellaril) and the antihistamine hydroxyzine (Atarax) did not help. The anticonvulsant carbamazepine (Tegretol) did nothing but give Heiner gum infections. Among the benzodiazepines tried, diazepam (Valium) 15 mg did not calm him or make him sleepy; neither did nitrazepam (Mogadon), which in higher doses led to vomiting. Among the neuroleptic drugs, thioridazine was only mildly calming; haloperidol (Haldol) and levomepromazine (methotrimeprazine) were stopped after 2 weeks because they not only made him more restless but he looked morose and seemed to be suffering. Sedatives such as barbiturates, and a combination of methaqualone plus diphenhydramine, did not help his insomnia.

The 1 medicine that really seemed to help Heiner was centrophenoxine (Lucidril). His cheeks turned rosy, he became happy and calmer, and he made eye contact for the first time. However, after only 3 days the doctor stopped the drug, saying that it was too dangerous. I did not mention Lucidril in the original article because I was not knowledgeable about the drug and could not find research on it.

None of the medicines we tried improved the severe problems that Heiner had falling asleep. I would have to lie in bed with him, holding him totally still for about 50 to 60 minutes, until he finally relaxed and fell asleep. Often he was back up at about 3 AM, coming into my bed. There he quickly fell back to sleep so that I could carry him back to his own bed. After a few years of this, his sleeping problems disappeared.

After the fruitlessness of our initial attempts at giving Heiner medicines, we gave up on all of them for a while. Some years later, the doctor prescribed the psychostimulant pyritinol (Encephabol), which Heiner took for many years. At first this drug appeared to have some benefit, but after a time it seemed to wear off. The broad failure of medicines led the psychiatrist to conclude that only pedagogical intervention would help. His advice led to the beginning of the behavior therapy described in this paper.

Initially, I wondered whether Heiner's autism could have resulted from his early digestive problems. In 1970, specialists at the pediatric clinic in Bern thought this possible but did not have any specific evidence. They diagnosed Heiner with a mild chronic intestinal infection, probably caused by a wheat allergy, but they did not find celiac disease. Biopsy showed a secondary sucrose intolerance. Later that year, Heiner's pediatrician put him on a special diet similar to one used for patients with celiac disease: no glutens, very little fat, and avoidance of vegetables known to cause digestive problems. Further, Heiner was not allowed any disaccharide sugars like sucrose or lactose; glucose was used instead. He was also given pancreatin (a combination of lipase, amylase, and protease). Used together for about 10 years (ages 4 to 13), these measures normalized Heiner's digestion. When he was 19, he starting living from Monday to Friday at a group home for adults with autism. The home served such poor and unbalanced meals that Heiner's digestive problems recurred and he had to be restarted on pancreatin. I finally stopped the drug in 2006 when it caused Heiner an outbreak of eczema. Since then he has not taken any medications or had digestive problems.

At age 3, Heiner began to have recurring tonsillitis. At age 3 years 6 months, he underwent a tonsillectomy and the infections stopped.]

[Once Heiner stopped speaking (at around 18 mo)], he remained mute except for an occasional “mama” and “papa” that seemed to have no real meaning. The dental sounds “t” and “n” that he had earlier were gone. [He did not hum or exhibit echolalia. He did not repeat syllables until these were trained, and he did not repeat phrases from songs unless these were specifically requested much later on.] At age 2 he tried to make himself understood by taking either my wife's hand or mine, and using it to point to the location where he wanted something. If the object or item was something he could get himself, he would do this instead. At about 4 years, we realized that his language comprehension was also very constrained. Only through systematic training, which occurred somewhat later, would this improve. Even before a systematic behavioral treatment plan was begun, when we tried to engage him Heiner would occasionally try to imitate the articulation locations of vowels, but he could not produce a sound. It seemed very unlikely that he would speak without some kind of training. His case showed us that engaging autistic children in any systematic way requires a strong, methodologically grounded therapy. The lack of this recognition among local speech professionals, together with most therapists' practices of bestowing gentleness, love, patience, and “trust in the healthy soul that has lost its way in a sick body,” really angered us, especially because we received absolutely no useful advice for years.

When Heiner was 5 years old, we took him to a Bern University speech disorders clinic in the hope that they would begin some kind of language therapy with him. However, they claimed that unless he could learn to concentrate and imitate, he could not be helped. Working with experienced teachers at a good kindergarten, he had made only small steps in this direction. My disappointment was that much greater because I knew from previous experience that there was only the smallest hope that he would learn language at this late age. The literature of the time claimed that a child who did not speak by age 6 years would never speak and would remain extremely mentally retarded. The fact that Heiner appeared to be an attentive and intelligent child was weak consolation, but it motivated us not to give up attempts at speech just yet. His intelligence per se could not be reliably tested. For example, his Vineland score at 5 years and 3 months gave him an intelligence quotient of 56, which put him at a social age of 2;11, an overly favorable estimation of his abilities.

Back to Top | Article Outline

Speech Sound (Phonemic) Training

The systematic speech training began when Heiner was 6 years and 5 months old. My wife and I began with a Hungarian “behavioral therapist” (Heilpädagogin) who had trained in Budapest and had her own clinic for many years. She had worked with mute and deaf subjects, but had little experience with autism. She did have experience with the Pavlovian method of behavioral conditioning, and taught us, as parents, a number of its principles. We had avoided speaking to Heiner in “motherese”—shortened, simplified words. We decided to train him in “high” German [rather than Swiss dialect German], for several reasons: Heiner's mother spoke high German, Heiner had few relationships with children who spoke the dialect, and articulation in high German is much more precise and can therefore be more easily practiced. Finally, early attempts at teaching Heiner to read and write high German had shown that this was not an impossibility [although, unfortunately, no further progress was made].

[Each week the Hungarian therapist gave Heiner at most 2 training sessions, each lasting 1 hour. Because she worked in her home, my wife or I had to take Heiner to her and we sat in on the sessions. In addition to her therapy, I worked with Heiner for about 6 hours a week, in 12 half-hour sessions. My wife and I became concerned about the therapist's rigor and, at times, well-meant fanaticism. Heiner would become aggressive when she held him tight and would not let him go, and because she did not give him a sense of “affectionate engagement” (“Zuwendung.” There is no equivalent word in English. Zuwendung is any behavior, verbal or nonverbal, that engenders affection in another.) Therefore, I increasingly took her place as the primary therapist. After about 18 months, we stopped seeing the therapist entirely, and I conducted Heiner's intensive training alone.]

The training [with me] seemed to be fun for Heiner, but only when he was in a good mood. [Early on, he would squeal when he did not want to work anymore and did not want to sit still. When he was upset, he could not even cry; he could only manage to pant.] The training time was gradually lengthened. It was accompanied by rhythm exercises and exercises involving body imitation. During the course of the work, I attempted to familiarize myself with behavioral therapy through readings (unfortunately, the possibility of undertaking speech training like that described by Lovaas et al3 was as yet unknown to me). At the same time, however, I maintained a critical distance from methods of operant language conditioning because of my language research knowledge and education in modern psycholinguistics and the theory of speech acts. [A speech act is a spoken or nonspoken communication, such as a greeting, request, invitation, compliment, apology, complaint, or refusal.]

There was a certain ethical mistrust of behavioral therapy among speech-language professionals in Switzerland at the time, yet I made myself familiar with what were then considered “authoritarian” training methods, realizing that these would be indispensable for a developmentally challenged child. I also had some practical experience in this method and with speech therapy, since I had tried it with my daughter to reduce her articulation impairments.

First, I determined that eye contact needed to be established (although I admit that the necessity of this came to me only later). I used operant conditioning with continuous reinforcement. At the table where the training was conducted, Heiner would receive a primary reinforcer for every eye contact made. In this case, it was a piece of salad, his favorite food. We could not give him sweets because of his special diet. Later he received glucose. He would also get praise and a fond rub on his back; these would later become his main reinforcers. The result was almost instantaneous. Soon, I reinforced eye contact only if it lasted at least 3 seconds, then 5 seconds, and later 10 seconds. I used the same conditioning method to teach him quiet sitting and imitating physical exercises. [At first, he was unable to sit quietly without some force, such as an adult's hands bearing down on his shoulders.]

The initial language therapy goals were to help Heiner perceive sounds auditorily in a meaningful way and to be able to distinguish individual sounds from one another. Upon presentation of a given sound, he was asked to imitate a given behavior associated with the sound. The first exercise was for him to knock on the table and be able to distinguish fast from slow knocking. At first his hands were led to do this, but this support was removed in a stepwise manner until he independently imitated the therapist's motions across from him. He was intermittently rewarded. The next step was for Heiner to learn to hit a xylophone, a tin drum, and a water glass as well as ring a bell. Again, we first used a hand-over-hand method. Then each of these objects was ordered according to its sound, which was then associated with the sound of 1 of the basic speech vowels, xylophone [a:], bell [e:], tin drum [o:], and water glass [i:]. First, Heiner learned to hit the appropriate object when he heard the correctly associated vowel. After a while, he was asked to say the vowels while he hit the objects. Soon he was able to hit the xylophone upon hearing the associated vowel, and then the tin drum and bell. The [i:] (water glass) took him longest to learn. Finally, the attempt to introduce a fifth sound [u:] failed. Heiner learned this last [u:] vowel only much later, once it was used together with the consonants [g] and [k] to distinguish it more clearly from [o:].

After he consistently responded to 3 vowels, Heiner had to learn not only to imitate the sound but also the articulation position of the vowel after observing the therapist's mouth. We worked together for a long time using a mirror. In this way, Heiner was able to combine auditory and visual impressions to produce consonants that were coupled with the vowels (ie, to produce syllables). Before training, he occasionally produced [p] and [m] by chance. These consonants were combined to form “pe-pe” and “me-me,” as well as “ap-ap” and “am-am.” After “pa-pa” he was required to distinguish “ba-ba.” After the bilabials, the dentals [t, d, n, l] were introduced. The back consonants were very difficult for him. Heiner produced the velars [g] and [k] as well as the glottal [h] successfully only by using a speech therapy technique. For example, we initially used a tongue depressor to ensure proper tongue placement for the [g]. Later, Heiner used his own finger in its place. Finally, it was enough for him to hold his index finger near his open mouth, at which point his tongue would retract automatically. New consonants were used and practiced in the context of syllables. At the same time, we introduced the difference between short and long, open and closed vowels, as well as umlauts and diphthongs. In addition to [i:] and [u:], the [y:] was difficult, as were [r] and the fricatives. When Heiner tried to vocalize an [r], it often made the vowels surrounding it sound wrong. Later he used [l] for [r]. The [j] was easier to learn, but the [v] was somewhat more difficult. He could produce [f] only weakly and when combined with other sounds, but was aided by blowing out a candle. The “ich” sound was weak at first, but surprisingly strong after exactly 2 years. Heiner did not learn the “ach” sound; this seemed to be replaced by the glottal stop [[Latin Letter Glottal Stop]]. At the same time, he was taught voiced and voiceless [z] and [s]. After 4 weeks of daily practice, using the same method as described for eye contact, he learned a clean voiceless [s] that was rewarded stepwise with the salad. We taught him the affricate “z” [ts] using a “trick,” by first training a strongly aspirated [t]. He could produce consonant clusters like [pl] and [gl] only once these sounds were made more distinguishable from one another by putting a [schwa] between the 2 consonant sounds and then blending the sound out again. The [ŋ] was trained by allowing Heiner to bind an alveolar [n] with a [g].

Back to Top | Article Outline

Training Short Phrases and Word Comprehension

The sound system or “babble training” was not much fun for Heiner. For this reason, we combined words with their meanings long before the whole sound training was finished. The training of individual phonemes was concurrent with learning the production of sentence-like requests such as “Bitte Salat” (please salad) and “Ich esse” (I am eating). It seemed reasonable to introduce 2- and 3-word phrases once Heiner could produce single words with reasonable, understandable articulation. The training of meaning began by our passively showing him objects and pictures in which these objects were found, while we spoke the objects' names. We asked him to point to either the object or picture when he heard the word (eg,“Where is the apple?”… “There!”). Pictures were easier to use at the table. Within the pictures, he was most likely to recognize objects that he knew from his environment. Specifically, his comprehension was most accurate for pictures that depicted the object in its most natural or realistic form (eg, in color and 3-dimensional). After some time, Heiner became much better at recognizing more abstract forms of objects, and he was able to recognize objects in reality that he had learned only from pictures.

[At first, Heiner confused such words as “tomato” with “banana,” and “donkey” with “horse.” I corrected such errors through clearly demonstrated speech in front of the mirror, and later while seated across from Heiner. I helped him by forming his mouth and tongue for correct articulation, first with his fingers and later also with a tongue depressor, as described above. When Heiner confused object names, I would set the objects or pictures in front of him and point to each one, saying the correct word. Then Heiner had to pick out the correct picture and was rewarded.]

After Heiner learned to imitate words on his own during the initial training, the type of question used in comprehension training was changed from “where” questions to “what” or “who” questions. Even if he could not quite articulate it accurately, he was asked to produce the correct word in response. At first he was asked to produce only 1-syllable words; soon thereafter, we expanded to 2-syllable to 4-syllable words. For combined word forms such as “teakettle” and “bathtub,” it was difficult for Heiner to deal with the doubled and hierarchical word stress. After some time, he was required to produce the associated article (eg, “der,” “die,” “das”). It was difficult at this point to determine accurately the degree to which what Heiner produced was merely imitation of syllable structures rather than speech production that referred to something substantive and meaningful. As Heiner's vocabulary increased, so did the incorrect agreement of the gender of the specific article and noun combination necessary in the German language. Heiner generalized and used a universal article sounding something like “de.”

At first we practiced only concrete nouns, because they often constitute the majority of productions in normal child language. Heiner learned only very few adjectives. We tried color terms first, but it was very difficult for Heiner to differentiate colors. In contrast, he learned the adjectives “warm,” “cold,” “broken,” “quiet,” and “good” (“good” in describing the taste of his food). Because he had broken many flowerpots while twirling them, the concept of “broken” was the clearest of all to him.

After presenting pictures of things to Heiner, we showed him pictures of actions. Now we asked him questions like, “What is the man doing?”…bathing, eating, etc., or “Who is that?” “What” questions were easier for him than “Who.” I practiced these with Heiner in concrete situations. I speculated, however, that he was more interested in the fact that someone was running or bathing than who the person was. For this reason, I decided to focus more on the naming of verbs and actions than objects and features. I purposefully attempted to set up most concepts with a focus on the verb, and from the verb to work on the building of the sentence. The fact that verbs tend to be less frequent in normal early child language was not a consideration at this point, but was probably also not an error since verbs may just not appear in the surface structure.

Another experience influenced the focus of the training and the nature of the practice in these dramatized situations. My wife and I had tried to force Heiner to speak by refusing to react to his attempts to point or take our hands to indicate what he wanted. Now Heiner began to grab our mouths and open them to show him what he should say when he could not. After about a year of teaching, Heiner produced his first 1-word sentence: an imperative speech act: “Geige!” (“violin,” meaning “Papa, get the violin!”). After a few months, he began to use other words as requests: “baden” (“bath,” meaning “I want a bath”) or “auto” (“car,” meaning “I want to ride in the car”), “brot” (“bread,” meaning “I want bread”), and “platte” (“record,” meaning “I want to hear the record”). [By this time, his squeals to indicate that he was finished with a training session were gradually being replaced by “aufstehen” (“stand up”).] The next step was that he had to say “please bath” or “please car.” In this way he learned to make his first syntactic connections. We arrived at this by modeling and speaking along with Heiner, followed by saying only the first syllable, then only the first sound, and then silently forming the first sound. In the end, we were able to eliminate these instructional aids entirely.

Back to Top | Article Outline

Training 2-Word and 3-Word Sentences

After 18 months of training, Heiner began to use 2-word sentences as they might appear in the adult language. Three months later, we instituted a systematic exercise of such sentences. I came up with a list of verbs that Heiner understood and that he could produce clearly. The verbs referred to actions that he knew from daily life and that interested him: eat, drink, sit, stand, lie, bathe, build, paint, come, cry, laugh, iron, sew, hop, clap, swing, drive, run, ride, whistle. These were all intransitive verbs (eg, “build” would be intransitive if there were no building built). These verbs could be part of fully grammatical 2-word sentences created from only a subject and a predicate. At the same time, I also constructed a list of nouns that could be combined with these verbs; mama, papa, Elke (Heiner's sister), Roland (Heiner's brother), the man, the woman, the boy, the girl, the doll, the car, the train. Some of the nouns required an article, so that now we could practice a 3-word sentence with a specific article. I presented 3-word sentences in analogical succession, working with pictures and photographs of the events. For example: “The man bathes, the woman bathes, etc.” “The man paints, the boy paints, etc.” “The man laughs, the woman laughs, the boy laughs, etc.”

Soon Heiner also learned to substitute “Bitte” (please) with “Ich will” (I want). Then, in place of “please,” he was asked to use “I want” (eg, “I want bread” or “I want stand up”—this phrase always followed working at the table). If he were asked to imitate a sentence like, “I want to have a record,” then he would shorten it to “I want record.” In this way he produced a 3-word sentence with a subsequent sentence part in which he treated the infinitive (to have) and the concrete noun (an accusative object, eg, record) in the same way. Next, the third-person singular and first person were used as subjects in 2-word sentences: “I bathe,” “I draw.” In this way, Heiner learned 3 inflected forms of the verb: “baden” (infinitive), “bade” (first person), and “badet” (third person). The usual difficulty with personal pronouns demonstrated in autistic children could be avoided through behavioral conditioning therapy. Heiner began to use the personal pronoun (first person) on his own. The second person was purposefully omitted until later in the therapy. The reversal of personal pronouns by autistic children is usually explained by echolalia or a disturbed sense of self. At the point of teaching pronouns to Heiner, he already seemed to have a clearly developed sense of Ich (“I”), potentially strengthened through the training. Amongst other evidence was his response of “ich” (“I”), and later “mir” (“to me”), when he was asked, “To whom does the bed (or “do the toys”) belong?,” although this response had never been explicitly trained. To the question of what he had done one afternoon, he answered several times, “Ich will badet” (“I want bathed”). This was a good sign to us that we could now introduce Heiner to the perfect verb form.

Back to Top | Article Outline


Whatever we may have achieved during the 2 years and 3 months of training Heiner, the fact remains that almost all of his production required many prompts (including questions and requests that specifically elicited speech acts). Many productions were still poorly articulated and Heiner did not reach a normal speech rate or intonation, that is, “sentence melody.” Judged against the accomplishments of a normal child's language, I could be tempted to conclude that we had achieved very little, but given Heiner's entirely mute state before training, the accomplishment was great. However, the question of whether Heiner's language development would accelerate beyond that point would not be answerable. At 8 years and 8 months, he entered a special boarding school that stressed perception abilities and cognitive, affective, and social development in the belief that these would improve communication. His success at this school was minimal. At the same time, I tried to continue training him at home on the weekends. We focused on improving articulation and sound vocabulary, introducing the perfect verb form, expanding syntax, and building his productive and receptive word knowledge. Altogether, semantic training was more important than syntactic instruction.

As the description of the instructional practice has shown, we refocused the original instruction with Heiner on operant conditioning, using basic principles developed by behavioral learning psychologists. From the beginning, we also included imitation in the learning process, to strengthen his imitative behaviors. It was only in this way that we were able to get Heiner to speak. To my knowledge, up to this point [1977] there are no other scientifically supported methods of language therapy with mute autistic children. Our instruction proceeded schematically from sounds to syllables, to words, to simple sentences. Conditioning and imitation are exactly what an exemplary representative of modern developmental psycholinguistics deems as not necessary for the natural acquisition of language.27 With Heiner, however, we found the behavioral conditioning method effective in the phonological, syntactic, and semantic domains. It would nevertheless be not only presumptuous but also totally in error to conclude, based on my experiences with Heiner, that the psycholinguistic controversies about language acquisition could be decided in favor of a stimulus-response model, and the critique by Chomsky, the founder of generative transformational grammar, against Skinner, invalidated.28,29 More important to me was my discovery that independent language acquisition by nonlanguage-impaired children must proceed quite differently in certain crucial ways. Heiner uses almost exclusively over-practiced speech patterns with a particular vocabulary, something that obviously does not happen in normal language communication. (It must be said, however, that this result was quite a victory for Heiner, even if he never veered from the learned pattern, since it was through the help of the ingrained patterns that a previously unknown method of making himself understood now became available.) Doubtless, Chomsky's current psycholinguistic theory of generative transformational grammar has invigorated language acquisition research with new and fruitful principles. The fact that conditioning methods lead to language acquisition in mute autistic children puts us in a better position to accept the possible influence of stimulus-response mechanisms in language acquisition by healthy children. Skinner's theory of verbal behavior, however, may be too simple to account for language acquisition. Instead, one might better adopt a model of representational mediation, such as that developed by Osgood.30,31

Although the benefits of behavioral conditioning in language acquisition were demonstrated clearly for me, I could not ignore the existence of language universals that result from inborn, biologically grounded human abilities. This was true also for the universals of language acquisition. For Heiner, the universal, inborn component was demonstrated most clearly in the phonological domain. As we worked on sounds, we probed many different routes of training, depending on which features Heiner could discriminate and which sounds he could produce. In doing so, I found that even in this case of language pathology, Jakobson's rules were upheld to a certain degree.32 In the beginning of the training, for example, Heiner already had [a], [p], and [m], those sounds that Jakobson mentions as being the minimal vowel and consonants found in normal speech. The consonants that Heiner learned could be learned only in the order that Jakobson claimed was “universal.” In contrast, Heiner deviated from normal acquisition of the production of vowel forms (something that had previously been shown in other children with autism). These had to be explicitly taught using behavioral methods. In Hamblin's language development program, as described by Wendeler,33 first the [a:], [e:], and [o:] are practiced. Before I was aware of this, I found that, before training, Heiner would produce only these vowels; the rounded [o:] was more difficult for him than the nonrounded vowels [e:] and [a:]. Production of the narrow-high vowels was also difficult, requiring a higher tongue position in the mouth ([i:], [ü:], [y:], [u:]). Most difficult was the combination of the phonological characteristics closed, round, and dark (the vowel [u:]). This difficulty was probably associated with the fact that Heiner had problems discriminating [i:] from [e:], and [u:] from [o:]. All of these difficulties, however, were surmountable through behavioral conditioning.

Certainly, one needs to be cautious in drawing general conclusions from 1 pathologic case. Yet after the experience with Heiner, I believe that neither the theories developed by the behavioral learning psychologists nor the psycholinguistic theories stemming from Chomsky's transformational grammar model are sufficient to explain language acquisition. Both attempts at explanation need to be part of a larger, overarching theory that would explain the relationships among the forms of sensorimotor, cognitive, affective, and social development that accompany language acquisition, and that would, in particular, explain the relationship between language and thought. Hypotheses regarding these relationships exist in the so-called area of “cognitive linguistics,” with the goal of a mental grammar,34 and even more in the psychology of language acquisition that developed out of the Genevan school of Jean Piaget. In the overarching theory it would be important, above all, to build on the notion of speech acts,35 whereby language acquisition could be understood in the clearest way by means of functional action relationships.

The most important result of our language training with Heiner was that he began to produce 1-word sentences on his own, he began to use the practiced set of more complex syntactic patterns as illocutionary and perlocutionary speech acts, and he thus began to use language functionally.

We tried to force Heiner to acquire functional language by denying him the things he wanted until he was moved from silence toward active speech. [For example, we asked him to say, “Ich will baden“ (“I want to bathe”) before allowing him to step into the bathtub.] Most probably, our experience points to the facts that language is acquired primarily in connection with actions and that psycholinguistic research and speech therapy must take better notice of the importance of the pragmatic side of language. I am referring to the very obvious observation that words or symbols referring to actions and events were more interesting to Heiner and easier to teach him than were words or symbols referring to people or objects. It was also clear that Heiner learned words for objects and people more quickly when these words were taught in combination with words for actions, than when the words were just presented repeatedly. In spite of these findings, I must also stress that it would have led to nothing if we had begun with the conviction that language consists always and from the start of speech acts and is acquired in normal language interactions (through the use of speech acts) or through language immersion. It would not have been possible to achieve Heiner's language independence had he been forced to speak in normal language contexts. Interaction had to be reduced to the simplest, extremely asymmetrical, and “artificial” form of training context. Heiner had to learn the phonological, syntactic, and semantic domains through conditioning before he could even partake in the interactive nature of speech. In the same way, we need to clarify and incorporate the notion of “affectionate engagement” [“Zuwendung”], a term that is used in the behavioral therapy and psychoanalytic literatures, so often with ambiguous or misunderstood meaning. With Heiner, we had to work on affectionate engagement through the conditioning of specific behaviors (mostly eye contact) and through my own behavior that helped to reinforce them (stroking his back, praise, etc.)

In the transition from single-word learning to the building of a simple syntax, I was left with the question of whether it was proper to use as the model for Heiner's 2-word subject-predicate sentences the developed child language and standard language of an adult, or perhaps instead to imitate early child-like constructions of 2-word sentences. Only a decision in favor of the fully developed syntax seemed logical to me—in the first place because, in spite of numerous studies, the syntax of young children has still not been investigated enough; second, because the observed forms can be explained in different ways; third, because there are obvious individual and mother-tongue-related differences; and, fourth, because for all these reasons it would not have been possible to determine the psychologically “simplest” 2-word constructions to use. Strengthening my decision was the fact that early child 2-word sentences always include a subject and predicate, with the possibility of later systematically adding further sentence constituents to these constructions. According to Bever,36 “subject-predicate” is one of the underlying relationships that children discover between the elements of language when the cognitive basis of language structures develops. This development mirrors itself again in the mental grammar of adults. According to Bever,36 most significant is the discovery of the differences between the subject-predicate and the predicate-object relationship. For this reason, my next step in training sentence patterns with Heiner had to lead to the construction of objects inside the verb phrase. (It happened first with the verb “wollen” {want}.) In other words, we trained sentences with verbs that take either an obligatory or an optional accusative object. McNeill27 describes the number of possible patterns in 2-word sentences, and their corresponding grammatical relationships. The most frequent patterns have verbs and nouns that stand in an object relationship to one another. Subject-predicate patterns occur much less frequently. Perhaps one can draw the conclusion from this that not only actions but goal-directed events are of particular importance for young children, so the speech act of “demand” stands in the foreground. For the language therapy with Heiner, I concluded that besides 2-word sentences with subject and predicate (with verbs that take one type of object), we should practice constructions with the word “bitte” (as in “bitte Brot”—“please bread”). This was also important because the implicit object relationship was very clearly the most salient in Heiner's spontaneous 1-word sentences. I also needed to consider whether we should be training so-called “pivot constructions.”37 The later version of pivot theory, which distinguishes between a central (closed) and an open class of words (Z-O constructions), would have to have been established first.38,39 But Z-O constructions make sense only if children produce them on their own, and many children hardly use them.39 In addition, these constructions appear to be incomplete surface reflections of deep underlying syntactic differences.27,40 Among Heiner's 2-word constructions, we could consider only those with “bitte” as being similar to Z-O constructions.

Since it is quite controversial whether language acquisition is affected more by the deep or the surface structure of the adult language (if the research into mental grammar doesn't make the belief in a deep structure superfluous!), there was luckily no other choice for me than to use the surface structure of the language for Heiner's instruction and just intuitively choose psychologically simple examples. In doing so, it was very helpful to refer to the classic research by Stern and Stern.41 The importance of the surface structure of the adult language is supported by the finding that autistic children with echolalia have a better chance of developing real language than do children who only comprehend language. The decision to use the adult forms of the language was also harder to abandon as Heiner started to be able to identify variants of utterances in the surface structure, and at times used these variants himself. The supported notion that language acquisition depends on a mental grammar slowly awoke in me the notion that the child's cognitive development must take priority in the teaching sequence; thus, it was senseless to keep conditioning syntactic patterns given Heiner's current level of cognitive abilities. In contrast, I did not want to abandon entirely the idea of fostering his cognitive development by conditioning specific sentence patterns through instructions regarding their functional use. Arguing in favor of continued syntactic training, one can consider the well-known phenomenon that children actually use certain language forms (eg, temporal or causal sentences) before they are truly able to understand the corresponding logical relationships. If language and thought depend on one another, as Vygotsky42 claimed, then improvement from usage to understanding should not be excluded. The Piagetian school, however, stresses that the foundation of cognitive operations lies not in the language but rather in preverbal sensorimotor actions. Cognitive development would therefore only be limited by language training.43 Because this problem remains unsolved, the dogmatic adoption of a particular model of language and cognitive development would be a mistake. Both methods should therefore continue to be tested, that is, through thinking to foster language and through speech practice to further drive forward the development of thinking.

Since I have already spoken in considerable detail about syntactic considerations, I must prevent a possible misunderstanding. My decision to use the developed adult language as the foundation for Heiner's practice did not by any means arise from a hypothesis about the priority or independence of syntax. Instead, the whole instructional practice became oriented toward semantic goals from the moment that Heiner began to produce sound combinations; that is, the child should be able to understand and produce meanings. The practice of syntax was the means to this goal, because I had been able to arouse little interest in Heiner by showing him and naming single items, and had so far not accomplished much in doing so.

Of the most commonly occurring constructions of young children's 2-word sentences,38 the 2 that I have emphasized with Heiner are the demand-wish and the describing of a simple event or situation. Heiner has not learned to ask questions. It will be a very important step to get him to do so. First, he must be conditioned for question sentences. It will not be crucial that he use the form of the question sentence. It is more about the ability, in addition to the “demand-wish” and the “description,” to produce a third important speech act: questions in the context of situations and event correlations. Heiner understands and answers questions, but he has not been motivated to ask them. [He often answers a stereotyped “yes” to questions, even when he means “no.” We take his “yes” to mean that he understands that he has been asked a question that requires an answer, but he still cannot respond appropriately.] Because of the importance of demands, I conclude that we should practice “where” questions first. He learned the demand speech act by himself, after he learned the corresponding sentence type through conditioning; thereby, the nonverbal could be converted into a speech-dependent interaction. It is still difficult to be sure whether each of his utterances about things and actions is a speech act of description, since these types of utterances occur exclusively in response to prompted questions. During instruction, at least, I could see the certain necessity that language develops not only through semiotic means as a symbol system, but also for understanding in an event-based corelational manner.

It follows from this report that by means of behavioral therapy methods, it is possible to build language, even in a child with severe autism, and even when the instruction begins relatively late in development. The severity of Heiner's impairment was apparent in all clarity when, at the age of 9¾, he could at last be tested (using Kaplan26), and his developmental level was shown equivalent to a child of 2.5 years. Through imitation lessons combined with language therapy, Heiner definitely learned to observe, and thereby achieved a better and more structured relationship with his surroundings. Through the learning of word meanings, a whole new world opened to him—or perhaps it was the other way around! In forming sentences, he was able to set up relationships between himself and his environment, and between things and people. Even if his language remains now and forever rudimentary, there is much that has been “won” in terms of interhuman understanding. One must caution that conditioning therapy may have achieved nothing more than a sort of language “dressage.” Yet in this artificial way, Heiner finally has a way of making himself understood. It is also difficult to unlock a language's sound system and sound relationships and to produce the syntactic relationships between those sound patterns; but this is a less fundamental aspect of the lesson. Only once a child actually produces its own speech acts can the language fulfill its task of facilitating an understanding with the surrounding world and thereby fulfill its social role.

[The most important factor in Heiner's acquiring language appeared to be our refusal to accommodate any of his nonverbal requests. Heiner learned to consider language less as a conditioned, reflexive production of sounds, words, and sentences, like a parrot, and more as a symbolic action, a specific act that could bring about a specific result.] The importance of this factor has been partially overlooked in previous reports of language initiation in autistic children, or the difficulty of the task has been underestimated. Many medical clinicians lack a basic understanding of the science of language. For example, Kehrer and Körber44 do not clearly describe the transition from conditioned speech forms to meaningful use of language. In a later overview article, Kehrer45 also leaves this most important problem open. It is not enough to conclude in a few words that trained echolalic speech must gradually be turned into a communication language through behavioral therapy. Language therapy is a highly complex task that must be conducted in a highly methodological but nondogmatic manner through a collaboration among psychologists, doctors, speech pathologists, behavioral therapists, and parents. The ongoing puzzle of language acquisition must be tied to the therapy. This work has the potential to achieve as significant findings as in aphasia research, findings that could further improve therapy.

Back to Top | Article Outline


Back to Top | Article Outline

Hellmut Thomke, translated by KB

Back to Top | Article Outline

1972 to 1974

Heiner's intense home treatment lasted for 2 years and 3 months, from age 6 to 8. Each week for the first 18 months, he had two 1-hour sessions with the behavioral therapist, with my wife and me looking on, and another 12 half-hour sessions just with me. For the final 9 months, I trained him alone, gradually lengthening the sessions and expanding their scope. My own job and other family obligations limited my available training time.

Back to Top | Article Outline


At age 8, Heiner went to a special weekday boarding school for cognitive development and language training. The school asked me to stop my home instruction because it would not be good to subject Heiner to 2 simultaneous training methods. I soon realized, however, that Heiner's language was regressing. So when he came home on weekends, I restarted my own method, sometimes sitting with Heiner as I had done before, but, more often, taking him on long walks. The length of time we spent walking would be hard to reconstruct. However, these walks gave us a natural setting for using language. Heiner learned to throw stones, run, sit on a bench, climb a tree, watch the trains go by, see a plane or a bird fly, pick berries. He also learned to differentiate large from small—stones, trees, etc. He was always required to describe the thing or action with words. Primarily because we could work together only on weekends, Heiner improved much less during this period than he had during the earlier, more concentrated training years.

Back to Top | Article Outline


At age 38, Heiner was doing quite well. He looked young for his age, almost child-like. He showed fewer and less bothersome overt signs of autism, such as hand flapping and whirling. Well socialized and calm, he behaved himself without problem on the street, in restaurants, and elsewhere in public. He had no sleep problems.

He still could not live alone. During the week, he lived and worked in a home for low-functioning adults with autism. He was at home with us every weekend and for about 4 week-long holidays a year.

Although Heiner could not live alone, his early training had given him some language independence. His active expressive vocabulary included about 470 words. His passive vocabulary, though much bigger, did not lend itself readily to tallying. Even though he had been trained only in high German, he understood words in Swiss dialect. He seemed to understand most things that happened day to day, like cooking, gardening, and shopping. He still liked music. He had also grown to enjoy listening to compact discs of dramatized fairy tales and children's stories with added music and sound effects. He could answer simple questions about the stories, like, “Which animals appear?” He spoke mostly in 1-word sentences when no one required more of him. We still had to prompt him even for some of his 1-word sentences. He used words spontaneously only when particularly fascinated by something, when the context absolutely required it, or when someone insisted on it. He still did not initiate much speech, and he tended only to whisper. Because his articulation remained poor, I continued to train him in pronunciation. He seemed to feel very disappointed when he tried to say something but could not articulate it properly.

At his weekday group home, Heiner liked using pictures to communicate. In 2004, the staff there had just begun trying to improve his communication with an electronic device similar to a DynaVox (DynaVox Mayer-Johnson, Pittsburgh, PA). The machine displayed icons depicting things and situations. Pressing on an icon made the machine speak the wish, for example, “I want to drink.”

That year we took Heiner back to the Bern University clinic and had him tested for nutritional allergies and for fragile X syndrome. All the tests were negative.

To summarize my thoughts about Heiner in 2004: I believe that his early training was responsible for his continued understanding and use of language. Our most important accomplishment had been improving his language comprehension. I found behavioral operant conditioning useful not only for language but also for decreasing Heiner's fears about things like going to the dentist. I wish that his intensive therapy could have started earlier and continued longer. Finally, I still felt disappointed by the lack of early support for conditioning methods when he was very young. Because of the “political milieu” among speech-language professionals in Switzerland, for several years they had emphasized facilitated communication almost exclusively.

Back to Top | Article Outline


At age 45, Heiner still lives in a group home. There he helps in the kitchen, with house chores, and in the garden. He can use tools like a saw, hammer, screwdriver, and drill. He is clean and orderly. He can take a bath once a week, ride, draw very simple pictures, and play simple music, as on a xylophone or drum. He behaves himself without fail. He has good rapport with the people who look after him, but less so with his housemates, most of whom have more intense disabilities than he and who speak not at all or very little. He still spends weekends at home with us, and is at home for longer vacations 4 times a year. He joins my wife and me when we go to a restaurant or attend a classical concert. At home he is very helpful, but needs rest breaks during which he pulls back to his room.

He can say roughly 500 words, although his “canned” utterances still require prompting. He continues to prefer to use pictures or gestures rather than words. I believe that as he ages, his understanding of language continues to improve.

Back to Top | Article Outline


The authors thank Hans Huber Verlag (Bern, Magazines) for permission to translate and republish Dr Thomke's paper.

This research was supported by the Cognitive Neurology Gift Fund and by the Therapeutic Cognitive Neuroscience endowment and gift funds.

HT: I thank Dr Boser for giving such great attention to the case of my son Heiner and for taking my 1977 paper, and carefully and most informedly translating it.

KB: My deep appreciation to Dr Thomke for allowing his son's story to be retold and for updating Heiner's progress. Thanks to my colleagues who inspired this project through their interest in learning how nonverbal autistic children have been helped to learn phrasal language after age 6: Erin Pickett, Olivia Pullara, Jessica O'Grady, and especially Barry Gordon, without whose perseverance this paper would not have been seen in print. Thanks to Sarah Wayland for advising on International Phonetic Alphabet notation. Thanks to Juane Heflin for reviewing the manuscript. Thanks to Edie Stern for her detailed and tireless efforts to prepare the paper for publication.

Back to Top | Article Outline


1. Thomke H. Sprachaufbau bei einem autistischen Kind: Überlegungen zum Spracherwerb und zur Sprachtherapie. [Language-construction in an autistic child: thoughts on language acquisition and speech therapy.] Schweiz Z Psychol Anwend. [Swiss Journal of Psychology and Its Applications.]. 1977;36:1–18
2. Brown JL. Prognosis from presenting symptoms of preschool children with atypical development. Am J Orthopsychiatry. 1960;30:382–390
3. Lovaas OI, Berberich JP, Perloff BF, et al. Acquisition of imitative speech by schizophrenic children. Science. 1966;151:705–707
4. Rimland B Infantile Autism: The Syndrome and its Implications for a Neural Theory of Behavior. 1964 New York Appleton-Century-Crofts
5. Pickett E, Pullara O, O'Grady J, et al. Speech acquisition in older nonverbal individuals with autism: a review of features, methods, and prognosis. Cogn Behav Neurol. 2009;22:1–21
6. Bourret J, Vollmer TR, Rapp JT. Evaluation of a vocal mand assessment and vocal mand training procedures. J Appl Behav Anal. 2004;37:129–143 quiz 143–144
7. Hingtgen JN, Coulter SK, Churchill DW. Intensive reinforcement of imitative behavior in mute autistic children. Arch Gen Psychiatry. 1967;17:36–43
8. Koegel RL, O'Dell M, Dunlap G. Producing speech use in nonverbal autistic children by reinforcing attempts. J Autism Dev Disord. 1988;18:525–538
9. Koegel RL, O'Dell MC, Koegel LK. A natural language teaching paradigm for nonverbal autistic children. J Autism Dev Disord. 1987;17:187–200
10. Martin R Out of Silence: A Journey into Language. 1994 New York H. Holt
11. Marshall NR, Hegrenes JR. Programmed communication therapy for autistic mentally retarded children. J Speech Hear Disord. 1970;35:70–83
12. Miller A, Miller EE. Cognitive-developmental training with elevated boards and sign language. J Autism Child Schizophr. 1973;3:65–85
13. Miranda-Linnè F. Individuals with autism spectrum disorders: teaching, language, and screening. Comprehensive summaries of Uppsala dissertations from the faculty of social sciences. Acta Univ Ups. 2001;103:1–96
14. Rutter M, Greenfeld D, Lockyer L. A five to fifteen year follow-up study of infantile psychosis. II: social and behavioural outcome. Br J Psychiatry. 1967;113:1183–1199
15. Rutter M, Rodefeld W Teaching Language to Autistic Children. 1974 Springfield, IL Department for Exceptional Children, Illinois Office of Education
16. Windsor J, Doyle SS, Siegel GM. Language acquisition after mutism: a longitudinal case study of autism. J Speech Hear Res. 1994;37:96–105
17. Welch MGTinbergen N, Tinbergen EA. Appendix I: retrieval from autism through mother child holding therapy. Autistic Children: New Hope for a Cure. 1983 London Allen & Unwin:322–326
18. Bondy AS, Frost LA. The PECS. Behavior. 2001;25:725–744
19. Newsom C, Rincover AMash EJ, Barkley RA. Autism. Treatment of Childhood Disorders. 1989 New York Guilford:286–346
20. Eldevik S, Hastings RP, Hughes JC, et al. Meta-analysis of early intensive behavioral intervention for children with autism. J Clin Child Adolesc Psychol. 2009;38:439–450
21. Heflin LJ, Simpson RL. Interventions for children and youth with autism: prudent choices in a world of exaggerated claims and empty promises. Focus Autism Other Dev Disabl. 1998;13:194–211
22. Stahmer A, Collings N, Palinkas L. Early intervention practices for children with autism: descriptions from community providers. Focus Autism Other Dev Disabl. 2005;20:66–79
23. Odom SL, Brown WH, Frey T, et al. Evidence-based practices for young children with autism: evidence from single-subject research design. Focus Autism Other Dev Disabl. 2003;18:176–181
24. Hess K, Morrier M, Heflin J, et al. Autism treatment survey: services received by children with autism spectrum disorders in public school classrooms. J Autism Dev Disord. 2008;38:961–971
25. Lenneberg EH Biological Foundations of Language. 1967 New York Wiley
26. Kaplan ELBühler H, Mühle G. Intonation und Spracherwerb. [Intonation and language acquisition]. Sprachentwicklungspsychologie. [Language acquisition psychology.]. 1974 Weinheim Beltz Verlag:152–174
27. McNeill D The Acquisition of Language. 1970 New York Harper & Row
28. Skinner BF Verbal Behavior. 1957 New York Appleton-Century-Crofts
29. Chomsky N. A review of B.F. Skinner's verbal behavior. Language. 1959;35:26–58
30. Osgood CHE, Sebeok TA Psycholinguistics: A Survey of Theory and Research Problems. 1954 Baltimore: Waverly Press
31. Hörmann H Psychologie der Sprache. [The psychology of language.]. 1970 2 Auflage Berlin Springer Verlag
32. Jakobson R Child Language: Aphasia and Phonological Universals. 1968 Paris Mouton
33. Wendeler JWing JK. Neuere Forschungsergebnisse. [New research findings.] Frühkindlicher Autismus: Klinische, pädagogische und soziale Aspekte. 1973 [Early child autism: clinical, pedagogic and social aspects.] Weinheim Beltz Verlag:283–308
34. Leuninger H, Miller MH, Müller F Linguistik und Psychologie. Ein Reader. Bd.I: Psycholinguistische Untersuchungen sprachlicher Performanz. Bd 2: Zur Psychologie der Sprachentwicklung. [Linguistics and psychology. A reader. Vol 1: Psycholinguistic investigations of language performance. Vol 2: On the psychology of language acquisition.]. 1974 Frankfurt a. M. Athenäum Fischer Taschenbuch Verlag
35. Searle JR Speech Acts. 1969 Cambridge Cambridge University Press
36. Bever TGHayes JR. The cognitive basis of linguistic structures. Cognition and the Development of Language. 1970 New York:270–362
37. Braine MDS. The ontogeny of English phrase structure: the first phase. Language. 1963;39:1–14
38. Slobin DFerguson C, Slobin D. Cognitive prerequisites for the development of grammar. Studies of Child Language Development. 1973 New York Holt, Rinehart & Winston
39. Leuninger H, Miller MH, Müller F Psycholinguistik: Ein Forschungsbericht. [Psycholinguistics: A research report.]. 1972 Frankfurt a. M. Athenäum Fischer Taschenbuch Verlag
40. Bloom L Language Development: Form and Function in Emerging Grammars. 1970 Cambridge, Mass MIT Press
41. Stern C, Stern W Die Kindersprache: Eine psychologische und sprachtheoretische Untersuchung. [Child language: A psychological and language theory-based investigation.]. 1907 Leipzig J.A. Barth Verlag (most recent reprint by Scientific Book Co., 1987)
42. Vygotsky L Thought and Language. 1962 Cambridge, MA MIT Press
43. Sinclair HHinde R, Stevenson-Hinde J. Some remarks on the Genevan point of view on learning with special reference to language learning. Constraints on Learning. 1973 New York Academic Press
44. Kehrer HE, Körber HP. Sprachbehandlung durch Verhaltenstherapie bei autistisch-mutistischen Kindern. [Language treatment through behavioral therapy with autistic-mute children.] Acta Paedopsychiatr. 1971;38:2–17
45. Kehrer HE. Verhaltenstherapie bei kindlichem Autismus. [Behavioral therapy in child autism.] Z Kinder Jugendpsychiatr. [Journal of Child and Adolescent Psychiatry.]. 1974;2:233–247

autism; conditioning; language acquisition; speech therapy; nonverbal

© 2011 Lippincott Williams & Wilkins, Inc.