Findings of a systematic review of interventions based on Gestalt Language Processing and Natural Language Acquisition

In this episode we chat with Professor Bronwyn Hemsley and Dr Lucy Bryant, both of the University of Technology Sydney, about their recent systematic review of interventions based on Gestalt Language Processing and Natural Language Acquisition. The discussion covers:

• What was the clinical question you were trying to answer?
• How did you go about undertaking the review? 
• What were the key findings?
• What does this mean for the speech and language therapy profession?

With thanks to panellists:

• Professor Bronwyn Hemsley, Head of Speech Pathology at The University of Technology Sydney and co-lead of the UTS Disability Research Network
• Dr Lucy Bryant, Senior Lecturer in Speech Pathology, The University of Technology, Sydney

Resources:

Bryant, L., Bowen, C., Grove, R. et al. Systematic Review of Interventions Based on Gestalt Language Processing and Natural Language Acquisition (GLP/NLA): Clinical Implications of Absence of Evidence and Cautions for Clinicians and Parents. Curr Dev Disord Rep 12, 2 (2024). https://doi.org/10.1007/s40474-024-00312-z
https://link.springer.com/article/10.1007/s40474-024-00312-z 

To read about the prevalence of echolalia mentioned in this podcast see the open access review: 

Sutherland, R., Bryant, L., Dray, J. & Robert's, J. (2024). Prevalence of Echolalia in Autism: A Rapid Review of Current Findings and a Journey Back to Historical Data. Current Developmental Disorders Reports, 11, 171–183 (2024). https://doi.org/10.1007/s40474-024-00311-0
https://link.springer.com/article/10.1007/s40474-024-00311-0

Please be aware that the views expressed are those of the guests and not the RCSLT.

Please do take a few moments to respond to our podcast survey: uk.surveymonkey.com/r/LG5HC3R

Transcript

Transcript Date: 

11 September 2025

Speaker Key: 

HOST:                         JACQUES STRAUSS

BH:                              PROFESSOR BRONWYN HEMSLEY

LB:                               DR LUCY BRYANT 

MUSIC PLAYS: 0:00:00-0:00:09

HOST:                      0:00:09 Welcome to the RCSLT Podcast. My name is Jacques Strauss. Gestalt Language Processing and Natural Language Acquisition, often called GLP/NLA, are approaches to language development that have become increasingly popular in speech and language therapy, including here in the UK, particularly with neurodiverse clients. Given the growing interest, it’s important that we share clear, evidence-based information on these approaches. 

Today, we are fortunate to be joined by Professor Bronwyn Hemsley and Dr Lucy Bryant, to talk about their recent systematic review, Interventions Based on Gestalt Language Processing and Natural Language Acquisition: Clinical Implications of Absence of Evidence and Cautions for Clinicians and Parents, published in December 2024. A link to this paper is in the show notes and is free to access.

                                 This conversation is intended to support RCSLT members in applying an evidence-based approach to practice, combining the best available research with clients’ values and preferences and their own clinical expertise, to provide effective and individualised support. And, as with any new intervention where evidence is not yet established, careful evaluation is key to make sure we deliver safe and effective services. As always, I started by asking our guests to introduce themselves, starting with Bronwyn.

BH:                           0:01:35 I’m a professor of speech pathology at the University of Technology, Syndey, and I’m head of discipline, and I am also the director of the Speech Pathology Clinic. I’m a clinical researcher and a speech and language pathologist in Australia. My main area of work has been in disability, particularly with minimally or non-speaking children and adults, and also in relation to assistive technologies and augmentative and alternative communication.

HOST:                      0:02:05 Thanks, Bronwyn. Lucy, I wonder if you could introduce yourself, please?

LB:                            0:02:08 Sure. I’m a certified practicing speech pathologist in Australia. I work at the University of Technology in Syndney. I’m a senior lecturer and a researcher and I’m also a clinician. I work in our campus clinic every week, primarily with children, with speech, language, fluency, all sorts of different communication needs. But I also teach in the classroom, primarily around working with children who have difficulties communicating, and language specifically is my area.

HOST:                      0:02:39 So, you conducted a systematic review to find any empirical studies that evaluate the effectiveness of GLP/NLA, but I wonder if first off, you could just give us a brief explanation of what GLP/NLA is.

BH:                           0:02:55 So, Gestalt Language Processing and Natural Language Acquisition, this was a framework developed by Marge Blanc in 2012 through the publication of a textbook about it, outlining a theory or ideas about how children, primarily those who are using echolalia or repeating back what they had heard, and particularly delayed echolalia, so after some time has passed, they are repeating back a phrase or a sentence or a word. So, Marge Blanc published a text outlining that these children go from the echolalia, from phrases to single words, whereas in language acquisition studies previous to this, people had followed children acquiring language from single words up to phrases and sentences, which is more familiar.

                                 So, this proposal that there were six stages meant that Marge Blanc also recommended certain strategies that should be followed in those stages of Natural Language Acquisition. And so, that involved responding positively and supportively around echolalia, so treating it as meaningful, which, it’s been proposed by other people as well, that’s not unique to GLP/NLA approaches, but certain things that Marge Blanc recommended as well, like avoiding single words until the child had gotten to a certain stage, stage four, or not using verb phrases, whereas in our usual practice we would do those things. 

There’s quite a bit of information in the textbook around those stages and what clinicians are to be doing, but there’s not really any research evidence underpinning those stages. So, that’s been where we were really interested to look at, well, what research is there that relates to GLP/NLA type approaches? Just so that we could better teach our students about the evidence base.

                                 While it was Marge Blanc who published in 2012, she and many others writing about it do cite earlier work of Ann Peters, as a linguist, who made some proposals around gestalt processing and made some observations of a very small number of autistic children who exhibited both analytic and gestalt processing styles. So, Marge Blanc herself often cites the work of Ann Peters.

                                 Another person is the work of Dr Barry Prizant, who took the idea a bit further from Ann Peters, and he suggested there were four stages for acquiring language in children using echolalia, again on a small number of children. But he importantly said there should be longitudinal research about this, and now Marge Blanc has claimed that she has done that longitudinal 15 years of research and presented it in her book. However, we haven't seen any report of the research in the book when we’ve looked at it. There are reports of children and different cases in the book, but it’s not what we would class as a research report. It’s not framed as a research study, it’s more framed as what is called a textbook.

HOST:                      0:06:24 And am I right in thinking that, despite the lack of research, some clinicians and parents strongly feel it’s an effective approach? Why do you think that is? 

BH:                           0:06:35 Well, there’s probably many reasons why it’s increasingly popular. It’s promoted very broadly on social media, almost universally positive stories are promoted. I’m not sure why that might be the case when there might be people who have not progressed with it, but we don’t tend to hear those stories, in social media at least. But we do know that people who criticise it are discouraged from doing so, or it’s not as safe to talk about what didn’t work than to talk positively about something working, so there’s that increased popularity.

                                 And I guess there’s the phenomenon that clinicians who say they think something works, they need to really know what was it that really made the child improve. Was it really what they did around GLP/NLA, or was it something else that they did that is common practice and that is recognised as improving language? At the moment, it’s not clear, because it’s not written down in a systematic way, what are people actually doing when they implement the therapy that leads to the improvements that they observe? 

A treatment study should be able to provide some measurements of that. It is interesting so many clinicians are saying, I’ve seen it work, I’ve seen it work in my clients, but why do we not yet see those accounts in a research report? One thing is, we don’t know how they’re identifying the children as GLP/NLA or requiring these interventions, so it’s hard to define who is the population and who isn't. 

And another thing is we know that the advice in the textbook includes some reasonable strategies we know could be effective. But the other thing that we’ve found out since doing a lot of workshops on our systematic review, in a lot of the workshops I asked clinicians, are you following the stages of the Natural Language Acquisition approach? And the vast majority of them are not actually implementing the stages, or following a staged approach, but they are implementing some of the strategies. Again, that’s the reason why we need systematic approaches, so that we can tease out, what is that’s working? If it is helping children, what is it, yeah, to make that conclusion? Because it could be that something else they are doing is really beneficial.

HOST:                      0:09:18 Can you walk us through the clinical or research questions that guided your review? What did you hope to determine about GLP in practice? 

BH:                           0:09:27 Well, in the absence of a systematic review, a new treatment that we’re teaching about in the classroom, we, if we can’t find a systematic review, then we will see if it’s worth doing one. And in this case, the popularity of GLP/NLA meant that our students were encountering it in their first day of placement, in their first year of study. We realised that it was really important to find any treatment studies that had been published or unpublished, that means somebody’s done a treatment study, they’ve written about it, but it’s not appearing in a journal yet. It’s in their cupboard and it’s somewhere, but it’s still knowledge. 

                                 We wanted to make sure that we knew if there were any treatment studies that looked at if children or adults improved in their communication, so that meant any method of communication, did it help them improve their language specifically? So, their ability to acquire language, develop language, use more complex language forms, or did it have any impact on their behaviour? And by behaviour, I’m not referring to social behaviour, I’m referring to behaviours of concern. Because if minimally or non-speaking children or adults do have behaviours of concern, we really do want to know if any kind of intervention is effective, because that way we can reduce their risk of being restricted in some other way. So, that was the aim of the systematic review, to find out were there any treatment studies, what were the effects of the intervention for those three things.

                                 Another important thing to remember is this intervention or this approach was really designed around those children or adults who are using echolalia as part of their communication. In the process of reading a lot of social media posts about GLP/NLA, it became clear to me that a lot of advocates for the method were suggesting that roughly 80% or 75% of autistic children are gestalt language processors, because this many are also using echolalia. And some people were posting that almost all autistic children are gestalt language processors. And so, I think it’s really important for people to understand more about research on the prevalence of echolalia. I think, Lucy, you can speak to that a bit more, given your recent review of echolalia prevalence.

LB:                            0:12:01 So, echolalia is a form of communication used by lots of different children, but it’s quite often associated with children who are autistic. It is the repetition of language that the child has heard. There’s two main types of echolalia. Immediate echolalia, where the child hears something said by another person or in the media, and repeats it immediately after they’ve heard it, and delayed echolalia, where they’ll hear it, they’ll store it in their memory, and they might repeat it again late, verbatim, they repeat it word-for-word what they’ve heard. It might be a line from a movie, it might be a line that they’ve heard from a family member or a friend, and so they echo or repeat it directly back.

                                 As Bronwyn said, in the GLP literature and the NLA literature, we see this common statistic, and it’s usually around 75-85% of autistic children echo, or even up to 100% of autistic children echo, and therefore, they are GLPs and we know that because they echo, it’s a bit of a circular logic sometimes. 

And so, the question was then, how do we know that? How do we know that that’s the prevalence of echolalia, that it is that common? What we wanted to do was look back at previously published research and literature to see, when we were looking, or when studies had looked at the characteristics of autistic children’s communication, how common was echolalia? And so, we looked back at all studies that were published since the latest diagnostic criteria, because the way that echolalia has been defined traditionally varies a lot. It’s really changed whether it’s communicative, or whether it’s a behaviour, over time, and there’s still a lot of debate about how that is adequately defined. 

And we found that there was a real lack of clear definition for what echolalia is within a research context, because clinic work and clinical education draws from this research so that can be problematic, because understanding what echolalia is is then something that is informed by very variable definitions. 

We went to the most recent definitions and the most recent diagnostic criteria. And of all the studies published since, or using that criteria, we found that within the sample, between 11 and 100% of autistic children echoed, and that was the prevalence. Anywhere from 11 to 100%, it’s a huge variation. And so, what that tells us is that the exact prevalence is really unknown. 

A lot of studies were then also citing research about what that statistic is, and we found that most recent research actually cites a study that comes all the way back from the 1950s. And it was based on institutionalised children, so its understanding of what autism is was very, very different to how we understand autism today. And in that study, it was actually only 36% of the autistic children who echoed, which was about a quarter of those who developed language. So, there was a lot who didn’t communicate, that were minimally or non-speaking, who didn’t echo, or didn’t have any communication to measure there. 

And the statistic had become quite warped over time, as one person had cited it and then another person had cited that person and another person had cited that person. And we ended up that the statistic that was largely cited in the GLP research was about seven or eight times removed from the original study. That really tells us that we can’t necessarily trust what this is telling us. And we don’t have that really clear of understanding of what is a GLP, if it’s based on purely echolalia and the understanding of echolalia itself is then based on research that is not current and is so highly variable.

HOST:                      0:16:06 Could you describe the methodology, the databases searched and the inclusion criteria, etc.?

LB:                            0:16:12 So, first of all, we published a protocol, which was an outline for what we were going to do in this study before we actually set out to do it. And by publishing a plan before we went ahead and did the systematic review, that keeps us accountable. Using terms about Gestalt Language Processing, we went looking through a lot of different sources, so there were 15 different scientific databases, including Google Scholar, which is accessible to everybody. 

We used publisher specific databases, so there was another four of those, there was three different clinical trial registers. We searched Google, we searched other websites, we searched for authors. We asked the world to send us anything that they knew of through communication boards and professional sources. And so, when we searched all of that, we brought back everything that we found that came back with the search terms around Gestalt Language Processing. 

And then we went through all of it. And when we went through it all, we were looking specifically for original research, so not reviews of previous papers, but the original research publication. We were looking for anything that talked about Blanc’s Natural Language Acquisition and Gestalt Language Processing, so citing that textbook from 2012. 

We were looking for things in English, this is a bit of a limitation of our review that we can go into. And we looked for treatment studies of any design. This is important, we weren’t looking for just randomised control trials or those really high-level studies, we were happy to find case reports and case studies with one participant, that’s a research study. 

And so, we were really happy to find anything from one participant to many, as long as there was a treatment that was implemented and there was measurements taken from that. And measurements might have been numbers, or they might have been descriptive, about the person’s experience and outcomes, it could have been that qualitative research too, so anything that didn’t meet those criteria was screened out. And everything was screened by two people, we had to agree on it to make sure that it was in. That way we were eliminating or reducing any bias that could come into this.

First, we looked at all the titles and abstracts, got rid of anything that wasn’t about GLP, and then we got all of the full texts together and read all of those too. And whittled it down until we got to our final result of any treatment studies of any design that were about Gestalt Language Processing, and that was none. We didn’t find anything, so out of our original 1,294 results that we found across all of those sources, we didn’t find a single treatment study that was about GLP or NLA.

HOST:                      0:19:08 Bronwyn, what does finding no studies tell us?

BH:                           0:19:15 Well, finding no studies is what’s called an empty review. And that means that even though we’d looked far and wide for any type of study, we went through all the things that we found and we came up emptyhanded, so even though we asked people to send us their studies, even if they weren’t published, we’ve got nothing. And certainly, since that time last year, we’ve had no more people saying we missed anything. 

So, we’re quite confident that it was an empty review, but a lot of people say, but what does an empty review tell you? And really what it tells us is that there’s no evidence in the research of the intervention being effective, but there’s also no evidence of it being ineffective. So, in a way, you can’t say it’s not effective, because there’s been no-one measuring properly with a research measure, with a research design, on it. 

Then what makes it really interesting in this field is how popular it is in the absence of research evidence. How do we, as a field, feel about adopting something so enthusiastically on the strength of testimonials? Given that there’s so many testimonials, almost every story you hear online is positive about it, really asks the question, why is there no research, and what would it take to do the research?

HOST:                      0:20:48 What do you think people who believe in the efficacy of GLP/NLA might say about your research?

BH:                           0:20:55 Well, I don’t have to really suppose about it, because I’ve seen the reactions. One of the reactions is, well, why are we only getting the negative criticism now? It’s been around since 2012, and, what, you're only criticising now in 2024? And my response to that would be, well, it wasn’t on my radar before the students started to say they were exposed in the clinic, and when we looked at the social media, I found the earliest posts around it were around 2021. 

So, 2020, 2021, was the time of COVID, lots of online learning, lots of social media connection, building up communities of practice online, and yet, it was hard to conduct research during COVID because of the restrictions on social distancing. So, probably, at the time, these two things happening, so that’s one reason, is that people have worried about, why are you only saying this now when it’s been going since 2012? And it’s really because it’s just come on people’s radar.

There’s so many podcasts about GLP/NLA, and here’s another one from us. The majority of podcasts are positive, overwhelmingly positive about it, so I think the growth in podcasting, the growth in self-publishing and online learning and now anyone can run an online workshop, which is fair, so, yeah, probably that’s one response from parents. 

Another one is that there’s some sort of conspiracy going on between us as reviewers, and to that we just simply say, that’s not true, we are simply researchers looking at the language evidence around autism. And not liking the result doesn’t mean that there’s some dark reason why we’re looking at the intervention. They’re the common problems.

But another really common question is, so, what do we do instead? Lots of people ask that. Another one is, what research needs to be done? If you're saying there’s been no research, why aren’t universities doing the research?

HOST:                      0:23:08 We will get to the research question in a moment, but first, were there any limitations in your review?

LB:                            0:23:14 The first, and I guess one of the most important limitations is that the research here is conducted by English speaking researchers. We searched databases that primarily database English language journals and English language research products, and we looked for articles published in English. Knowing that the GLP/NLA movement is a little bit more global, we may have missed things that are published outside of the English language, we could have missed things. We haven't been informed of anything, as Bronwyn said, from anybody since we published the review, but that’s a notable thing.

                                 And we were also only looking for treatment studies, and so people have pointed out that we haven't necessarily covered other anecdotal reports and other forms of publication outside of treatment evidence. And that is a limitation of a systematic review, it does only look at treatment evidence. And so, to address that particular limitation, we are expanding our search and we are looking for other forms of publication around Gestalt Language Processing and Natural Language Acquisition. And that’s another research study and perhaps something to discuss at a later point, where we hopefully will be able to get a better view of what is out there and what is published beyond that treatment research.

HOST:                      0:24:39 Given your findings, what kind of research would you most like to see?

BH:                           0:24:44 Well, ideally, systematic review would include a number of different designs, and then be able to take the results from each of those and analyse, okay, so what do all those different designs tell us? There’s the obvious single case, multiple baseline design, which is the most feasible in this instance, because you really, you need to make repeated measures of our outcomes along a baseline and then throughout a treatment phase and then after in a follow-up. It’s not highly complex to do, it’s something that’s very common in the field of AAC.

                                 Now, if you go on from there and you’ve got small group studies with some control, so there’s a comparison group who have a different treatment or have the treatment after and they swap the groups, those sort of group studies. Then there’s randomised controlled trials, which involve groups of children who are similar and comparable in different aspects that they can be compared, and then randomising them to a treatment group and a control group. That’s something that wouldn’t straightaway happen. Usually researchers would start with something small-scale, to try and approve the concept or run a pilot study. We would have been happy with any of those designs in a treatment study.

                                 But what I would say is that clinicians who are reporting it’s helping their clients, in the absence of research evidence, clinicians have an ethical responsibility to be collecting data and doing that systematically, so that they can track what is the child doing before they start the intervention, and then tracking those outcomes during the treatment and following up afterwards. Clinicians can also do their own clinical research in their populations, whether they do that as a formal research study with ethical approval or whether they’re doing it as part of their clinical documentation.

HOST:                      0:26:53 And this is a good time to remind listeners about ROOT, the RCSLT Online Outcome Tool, which is a fantastic way to gather clinical evidence using the TOMs scales for this kind of research. Now, I guess the question then is, what are the clinical implications of this review?

BH:                           0:27:12 Sometimes it’s important to go back to the clinical recommendations, if you like, for this group, that they do treat echolalia as communication, that they do respond to children positively and strengths-based approach, a person-centred approach, and all of that, which is part of the guidance in working in speech and language therapy for any client. 

                                 But the other thing would be, make sure to assess the child’s abilities and skills across their range, for expressive and receptive language. We’re starting to see people saying, well, because they’re a GLP they won't follow questions, they won't understand questions, and also, they wouldn’t manage an assessment of their receptive language. It’s really important that they do have access, they have a right to having those skills assessed, as any child does, so just offering the full range of services to the client would be another recommendation. Whether it’s a formal or informal or a non-standard or a standardised assessment, that would be the clinician’s decision as to how they do the assessment.

                                 Because clinicians have autonomy, they can take the results of our systematic review and take it into account. It doesn’t mean that they have to do or not do certain things, it just means they have to take it in. That means they should advise parents on the absence of research evidence, so that parents can make an informed decision that they are going into an intervention without it having research evidence. 

                                 And then if you're thinking about this population, there’s all sorts of recommendations coming through that also are not evidence-based to do with the design of their AAC or their augmentative communication system, such as, using a lot of phrases in their communication board or in their speech device. And to that, I would just remind people that children need multiple ways to communicate, high technology, low technology, unaided communication, with facial expressions, gestures, all of that. 

All of that’s important, but it’s important not to have just a sole focus on phrases. Even if the child’s using echolalia, they probably have the potential to learn single words. Find out what is their symbolic understanding and then teach them around those meanings of symbols, so that they’re not only limited to stock phrases or phrases or echolalia or whole sentences when they might be needing to use single words or put two single words together.

LB:                            0:29:54 We know that there are commonalities that pop up in many GLP recommendations that are also common across other evidence-based approaches to language intervention. And they include things like modelling language for a child, narrating your day, talking to the child about what is happening in their environment around them, giving language to everyday actions and everyday interactions, listening to the child, responding when they talk. Responsivity is a really important part of developing that knowledge that communication is a two-way, back and forth interaction. 

And so, when a child makes any attempts to communicate, giving something back, acknowledging that, whether it is just repeating what they have said, adding to it, acknowledging it with interest, oh, yes, that was so interesting, tell me more. Even if you don’t understand it, it gives that sense of communication and interaction. We know that responsiveness is a really important thing that parents can do with any child, regardless of how their language development might be defined, and modelling that language for them and showing them what language can be acquired to talk about the things in our environment. 

There are some really important elements to that that are not recommended as part of GLP recommendations and supports, that includes things like using verbs or actions words. GLP/NLA approaches suggests leaving those out until a child is actually starting to put words together. But what we actually know from research is that given that especially in English a verb is the central and necessary component to make a sentence, we don’t model verbs and action words for our children. They don’t have that key to put sentences together, so having a strong knowledge of action words is actually predictive of how well a child will be able to use grammar in the future.

HOST:                      0:31:54 A very big thank you to Bronwyn and Lucy for their time today. As always, please see the show notes for a link to the article and other resources of interest. Given the diverse views around Gestalt Language Processing and Natural Language Acquisition, it is important for clinicians to remain informed and look out for any new journal articles or studies investigating this approach. And when we do learn about interesting material, here at RCSLT we’ll be sure to discuss them on the podcast and other channels.

                                 Please do rate and share this podcast with your colleagues, so that we can continue to advocate for SLTs, service users and their families in the UK and further afield. Until next time, keep well.

MUSIC PLAYS: 0:32:33-0:32:44 

END OF TRANSCRIPT: 0:32:44