by Hannah Blackwell, M.A., CCC-SLP
Speech-Language Pathologist
Duncan Lake Speech Therapy
At Duncan Lake Speech Therapy, we highly value patient autonomy, client-centered care, and neuroaffirming practices. These values guide our therapy practices, and our clinicians strive to stay up to date with best practices and emerging evidence. Recently, an article titled ‘A Response to Blanc and Colleagues’ Viewpoint on Gestalt Language Processing and the Natural Language Acquisition Protocol: Concerns and Common Ground’ was published, offering commentary on a relatively new approach to language treatment for autistic children. Because all clinicians at Duncan Lake Speech Therapy are trained in Gestalt Language Processing (GLP) and the Natural Language Acquisition (NLA) protocol, we felt it was important to share how we approach GLP/NLA in light of the questions raised in this article.
What is Gestalt Language Processing (GLP) and Natural Language Acquisition (NLA)?
The way humans interpret, comprehend, and use language is encompassed in the term language processing . Typically developing children often follow what’s called “analytic language processing” (ALP) as they acquire language skills. ALP language development typically starts with assigning a single meaning to a small unit of language (i.e., a single word) and progresses to building length of utterances by combining words, until conversation-level language is reached (Blanc et al., 2023). Research into language processing has revealed that there may be another form of language processing called “gestalt language processing” (GLP). GLP begins with larger single units of language (i.e., full phrases/sentences) in which a single meaning is assigned to a larger “chunk” of language which are referred to as gestalts (Blanc et al., 2023; Peters, 1977). This type of language processing then progresses to breaking down the larger units into smaller units until a single meaning is assigned to a single word, before following a similar trajectory to ALP and combining single words into unique, meaningful utterances.
Where ALP tends to have more of a linear trajectory, GLP does not. Children may move between stages of language development fluidly as they progress toward language flexibility. Clinicians often observe a wide range of language processing patterns, and many describe this as a continuum, with some able to develop language skills more easily, and others having difficulty developing language skills without intervention. And, of course, there are others who fall somewhere in the middle. Both forms of language processing lead to flexible, self-generated speech, and while some gestalt language processors are able to move through the phases of language development without intervention, some have more difficulty doing so.
Therefore, the Natural Language Acquisition (NLA) protocol was developed by Marge Blanc to help those who fall more toward the “fully gestalt” side of the spectrum. This approach to therapy breaks down GLP into 6 stages and provides a framework for SLPs to follow, including 6 stages of GLP, how to model language within each stage, and how to implement child-led, naturalistic therapy techniques. It is important to note that no one can definitively be placed in one box or another simply based on the presence or lack of diagnosis. The NLA protocol identifies that autistic gestalt language processors are often the most obvious simply because they are often the most delayed, but not ALL gestalt language processors are autistic, just like not all autistic people are gestalt language processors. Because GLP/NLA is relatively new, there is not much research completed on it and it makes sense that people would question its validity.
This brings us to a recent article that was published which identified 5 major concerns with GLP/NLA: (1) GLP is contradicted by existing research evidence, (2) claims about language processing style are based on only expressive (spoken) language, (3) it assumes ASD kids don’t understand individual words in their echolalia, (4) it recommends unfounded (potentially harmful) language modeling strategies, (5) it is not supported by peer-reviewed research (Lorang et al., 2025). While we cannot conduct research on GLP/NLA ourselves, we can address these concerns from our perspective in terms of our own clinical practice.
Concern 1: The idea of GLP is contradicted by existing research on how autistic children process language
The first concern brought up by Lorang et al. (2025) is that the idea of GLP is contradicted by existing research, writing “a key idea within GLP is that many autistic children process spoken utterances as ‘chunks’ or as ‘an unanalyzed whole, that is, they are not aware that a language gestalt could be made up of individual words that are combined’ (Blanc et al., 2023, p. 1280; see also Blanc, 2012; Peters, 1977).” They continue in their argument by explaining that studies that have collectively included 300 kids have shown that ASD children age 3-17 years suggest that ASD children do not process spoken utterances as chunks (Bavin et al., 2014, 2016; Brennan et al., 2019; Brock et al., 2008; Hahn et al., 2015; Pomper et al., 2021; Prescott et al., 2022; Venker et al., 2013, 2019; Zhou et al., 2019). When looking at the Blanc et al. (2023) article from which the quote referenced by Lorang et al. (2025) was taken, in this quote, the authors were referring to a specific study conducted by Peters (1977) in which she observed children who were at the single-unit stage of development and was describing the specific individuals in her study, with Blanc et al. (2023) saying that Peters “observed that young children who are at the single-unit stage developmentally seem to perceive each unit as an unanalyzed whole, that is, they are not aware that a language gestalt could be made up of individual words that are combined.”
While a single research article is not enough to prove or disprove a concept, it does show that at least one study documented patterns that share similarities with ideas later described in GLP, though this is far from conclusive. Research directly examining GLP and the effectiveness of NLA is still quite limited. Therefore, while there is research showing that some ASD children do process single words as single units of language, these findings may not capture the full range of language development profiles we observe clinically in autistic and non-autistic children. As discussed, a diagnosis of autism does not automatically mean a child is a gestalt language processor, and likewise, a child without an autism diagnosis is not necessarily an analytical language processor. There is no available research focusing solely on those children who are suspected to be gestalt language processors. Therefore, this concern reflects a broader challenge in the speech-language pathology field: that since there is such limited research available regarding GLP, more research is needed. This rings true not just for GLP, but across many areas of language development.
Concern 2: The conceptual premise of GLP/NLA is Fundamentally Flawed
This concern centers on the distinction between expressive (spoken) and receptive (understood) language, which are separate concepts and one should not be used to determine the level of the other. Because GLP proposes that production of delayed echolalia (expressive language) indicates a GLP style of comprehension (receptive language), the authors were concerned. While this is a trend that has been observed in the research conducted by Blanc et al. (2023), Lorang et al. (2025) does bring up a valid point: we cannot assume receptive language abilities based on expressive language. This is particularly relevant for children who use AAC or have limited spoken language, as their receptive capabilities cannot be determined based on their expressive language skills. Our SLPs are well aware of this fact, and assess expressive and receptive language separately in all patients brought into our clinic. It is best practice to use multiple modes of assessment to determine a child’s areas of need, and that is no different if GLP is suspected in a child.
The language sampling approach suggested by the NLA protocol is used in conjunction with formal and/or informal testing to assess expressive and receptive language abilities. All pieces of the assessment are then used to determine what a child needs to target to improve their language skills. Just like children who are analytical language processors, receptive language skills can be higher or lower than expressive language skills and vice versa; therefore, a child who has higher receptive language skills could still benefit from the NLA framework if they are using mostly echolalia to communicate. Additionally, as with any therapeutic approach, the plan of care, regardless of the therapeutic framework followed, is personalized to the client’s specific language profile.
Concern 3: GLP/NLA assumes autistic children do not understand individual words in their delayed echoes, which presumes incompetence and is inconsistent with neurodiversity
The authors’ main concerns in this section is that by engaging in NLA protocol with an autistic individual, clinicians may assume lower cognitive functioning than what is present and may withhold exposure to specific types of language concepts (ex. math, literacy, verbs), which could have serious consequences later in life. They interpret this assumption as presuming that the client is incompetent and that this is inconsistent with neurodiversity. As a clinic, we agree that any assumption of lower cognition due to diagnosis or language functioning is non-neuroaffirming. The choice to use NLA or consider GLP does not, in itself, reflect any assumptions about a client’s abilities. It is the clinician’s intentional, neuroaffirming implementation that guides how the approach supports the child.
With any therapeutic approach, the therapist must personalize it to the client; no approach is one-size-fits-all. When implementing any therapeutic approach, including GLP/NLA, it is our responsibility as clinicians to presume competence and ensure that our implementation reflects that. We must use the evidence we’ve gathered to develop a personalized plan of care. With any patient, however, there are often a multitude of targets we could choose to work on. Therefore, we must prioritize what we will target and measure progress with based on each patient’s unique language profile. This does not mean that we will cut out or discourage the other parts of language that children have difficulty with. In therapy sessions, we implement all therapy via a patient-led approach, meaning we take the client’s lead in therapeutic activities and work their goals into their chosen tasks. We pick targets to work on and focus on those in therapy, but we do not discourage or avoid modeling language that is not included in the chosen language targets. For example, if we choose to target observable concepts (noun, adjective, location) we do not ignore if the child produces other word forms (ex., verbs, prepositions, numbers), rather we use the language they produce to continue incorporating the concepts we’ve chosen as their goal targets.
Concern 4: The NLA protocol’s stage-based language modeling recommendations are unfounded and may be harmful
This concern reflects the broader issue that research on GLP/NLA is still emerging and incomplete, as well as the fact that some NLA strategies differ from those typically used with analytic language processors. As stated previously, it is well-recognized in the SLP field that many areas of language development still need further research. While the research used to support GLP/NLA is preliminary, it only points to the need for further research to discover more about this concept. Clinical experience can help guide decisions when research is limited, as long as those decisions are continually monitored and individualized, which is the standard we hold ourselves to in our clinic.
One indicator discussed in NLA training (Meaningful Speech) is that some children who do not make progress with traditional analytic approaches may benefit from GLP-informed strategies. Therefore, part of our assessment and therapeutic process is first trying the well-established language therapy approaches prior to attempting NLA to allow us to better differentiate between ALP and GLP. Concern 5: GLP/NLA is not supported by peer-reviewed, empirical research evidence A common theme both in Lorang et al.’s (2025) article as well as in this blog post: there is simply not enough research. Although the research base is still limited, early observational studies on delayed echolalia—including the ones cited by Blanc (2023)—did describe language patterns that resemble what we see clinically in some children, which is part of why the NLA framework was developed and why we continue to find it helpful in practice.
Also…
In the SLP field, we use the Evidence-Based Practice (EBP) framework when choosing treatment plans. EBP is comprised of three branches: (1) the best available research evidence, (2) the clinician’s clinical expertise, and (3) the patient’s values, preferences, and circumstances. While peer-reviewed, empirical research is the golden standard of research, if that is not available, we can use what’s available in the research to inform our clinical decision making. Combined with research, though, we must also use our own personal, clinical expertise and experiences with our patients as well as our patient’s personal circumstances to determine the best treatment plan.
At Duncan Lake Speech Therapy, we use EBP in our clinical decision making, and we have found the NLA approach to be helpful for many of our patients. As with any approach, it is not appropriate for every patient that comes in, but we take care to evaluate each patient individually and choose the approach that will benefit them most, whether that is NLA, typical language learning strategies, or something else altogether.
In closing, the concerns brought up in this article highlight important questions and areas where more research is needed, but they do not prevent us from using GLP/NLA thoughtfully and responsibly when it aligns with a client’s needs. This article contributes valuable questions that can guide further research, and we hope to see that research continue so we can refine our clinical decision-making to best support any client who walks through our doors. Importantly, the article also reinforces the continued use of general treatment strategies that are foundational to effective therapy: establishing trust and connection with children, collaborating with families, following the child’s lead, honoring all communication modes (including echolalia), and listening to autistic voices. In the meantime, we will continue using our clinical judgment, our clients’ circumstances, the research that is available, and all appropriate therapeutic resources to ensure our clients make meaningful progress in developing their communication skills.
Resources:
● Bavin, E. L., Kidd, E., Prendergast, L., Baker, E., Dissanayake, C., & Prior, M. (2014). Severity of autism is related to children’s language processing. Autism Research, 7 (6), 687–694.
● Bavin, E. L., Prendergast, L. A., Kidd, E., Baker, E., & Dissanayake, C. (2016). Online processing of sentences containing noun modification in young children with high-functioning autism. International Journal of Language & Communication Disorders, 51 (2), 137–147.
● Blanc, M., Blackwell, A., & Elias, P. (2023). Using the Natural Language Acquisition protocol to support gestalt language development. Perspectives of the ASHA Special Interest Groups, 8 (6), 1279–1286.
● Brennan, J. R., Lajiness-O’Neill, R., Bowyer, S., Kovelman, I., & Hale, J. T. (2019). Predictive sentence comprehension during story-listening in autism spectrum disorder. Language, Cognition and Neuroscience, 34 (4), 428–439.
● Brock, J., Norbury, C., Einav, S., & Nation, K. (2008). Do individuals with autism process words in context? Evidence from language-mediated eye-movements. Cognition, 108 (3), 896–904.
● Hahn, N., Snedeker, J., & Rabagliati, H. (2015). Rapid linguistic ambiguity resolution in young children with autism spectrum disorder: Eye tracking evidence for the limits of weak central coherence. Autism Research, 8 (6), 717–726.
● Lorang, E., Mathée-Scott, J., Johnson, J., & Venker, C. E. (2025). A response to Blanc and colleagues’ viewpoint on gestalt language processing and the Natural Language Acquisition protocol: Concerns and common ground. Perspectives of the ASHA Special Interest Groups , 1–8. https://doi.org/10.1044/2025_persp-25-00055
● Peters, A. M. (1977). Language learning strategies: Does the whole equal the sum of the parts? Language, 53 (3), 560–573.
● Pomper, R., Ellis Weismer, S., Saffran, J., & Edwards, J. (2021). Coarticulation facilitates lexical processing for toddlers with autism. Cognition, 214 , Article 104799.
● Prescott, K. E., Mathée-Scott, J., Reuter, T., Edwards, J., Saffran, J., & Ellis Weismer, S. (2022). Predictive language processing in young autistic children. Autism Research, 15 (5), 892–903.
● Stiegler, L. N. (2015). Examining the echolalia literature: Where do speech-language pathologists stand? American Journal of Speech-Language Pathology, 24 , 750–762.
● Venker, C. E., Eernisse, E. R., Saffran, J. R., & Ellis Weismer, S. (2013). Individual differences in the real-time comprehension of children with ASD. Autism Research, 6 (5), 417–432.
● Venker, C. E., Edwards, J., Saffran, J. R., & Ellis Weismer, S. (2019). Thinking ahead: Incremental language processing is associated with receptive language abilities in preschoolers with autism spectrum disorder. Journal of Autism and Developmental Disorders, 49 (3), 1011–1023.
● Zhou, P., Zhan, L., & Ma, H. (2019). Predictive language processing in preschool children with autism spectrum disorder: An eye-tracking study. Journal of Psycholinguistic Research, 48 (2), 431–452.
Recent Comments