Help for translation

59 views
Skip to first unread message

Sophie Gerday

unread,
Jun 6, 2021, 4:08:11 PM6/6/21
to FluencyBank

Hello all,

We are transcribing play sessions and we are needing some recommendation to save us in our transcriptions. When non-stuttering child says: 

« he’s m… he’s making noise”, how do we transcribe the sentence?

*CHI:    <he’s &+m> [/] he’s making noise      Flucalc counts 2 TD (1 phonological fragment and phrase repetition) et 0 SLD.

OR
*CHI :   <he’s m(aking)> [/] he’s making noise     Flucalc counts 1 TD (phrase repetition) and 0 SLD.

My apologies for this basic question but I’m a new user of CLAN/CHAT.

Thank you very much for your help!

Sophie

Nan Bernstein Ratner

unread,
Jun 6, 2021, 4:18:26 PM6/6/21
to Sophie Gerday, FluencyBank
There wouldn't be a hard and fast rule, but the second option (deleting most of the word) seems odd - it's mainly used for rule-governed omissions, such as (a)bout or swimmin(g). I'd go with the first. I presume that the behavior also lacks the tension of a stutter, since you are transcribing behavior, not representing a diagnosis. If it had any stuttering qualities, regardless of how you'd describe the child, I'd go with stuttering notation if it were a tense repetition. 

best,
N

Nan Bernstein Ratner, F-, H-ASHA, F-AAAS, ABCLD
Professor
Hearing and Speech Sciences
University of Maryland
0100 Lefrak Hall
College Park, MD 20742


President, International Fluency Association (IFA; http://theifa.org)
Director, University of Maryland Autism Research Consortium (UMARC), www.autism.umd.edu
Faculty, Language Science (languagescience.umd.edu; Neuroscience & Cognitive Neuroscience (NACS, nacs.umd.edu), Developmental Science Field Committee




--
You received this message because you are subscribed to the Google Groups "FluencyBank" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fluencybank...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/fluencybank/c77b3e7a-f26d-434c-ad46-4719adb7e26dn%40googlegroups.com.

Carol Hubbard Seery

unread,
Jun 6, 2021, 5:43:02 PM6/6/21
to Nan Bernstein Ratner, Sophie Gerday, FluencyBank
Hi Sophie, 
    Speaking from experience with the Illinois Stuttering Research group, we interpreted the disfluent behavior you described only as a phrase repetition, because the unit being repeated had crossed word boundaries (thus was a greater sized segment repeated than a word repetition). It would have been considered similar to a multisyllable word/segment repetition which were grouped together with the phrase repetitions as longer unit/segments being repeated).  Another category that could make sense is 'unfinished word/phrase' and revision, also because of the nature of disfluency. 
     I am not familiar with the 'phonological fragment' category so I cannot speak to whether that applies or not.  However, I will say that we would not have classified the disfluent behavior with any category that might be considered stuttering-like/core/primary/within-word because, like Nan, I assume that there was no perceived tension at the moment of the break in the forward movement of speech, especially as a non-stuttering speaker, you said. 
      I hope this explanation makes sense.  
         Carol
Carol H. Seery, Ph.D., CCC-SLP, Associate Professor Emeritus
UW-Milwaukee Dept of Communication Sciences and Disorders
Title use:  Dr. Seery     Pronoun use: she/her/hers
EMail:  cse...@uwm.edu    Work phone no.: (414) 229-4291 (thru MS TEAMS)


From: fluen...@googlegroups.com <fluen...@googlegroups.com> on behalf of Nan Bernstein Ratner <nra...@umd.edu>
Sent: Sunday, June 6, 2021 3:18 PM
To: Sophie Gerday <fofi...@gmail.com>
Cc: FluencyBank <fluen...@googlegroups.com>
Subject: Re: Help for translation
 

Yairi, Ehud

unread,
Jun 6, 2021, 6:18:27 PM6/6/21
to Sophie Gerday, FluencyBank, Carol Hubbard Seery, Nan Bernstein Ratner
I do not know how this arrived in my mailbox and I have not done this kind of work for a good number of years.  Still, my immediate reaction was "phrase repetition" which, most typically, is not perceived as stuttering. We classified this type with the "Other Disfluencies" group.

Ehud

Ehud Yairi
Professor Emeritus
University of Illinois, Speech & Hearing Science
Tel Aviv University, Communication Disorders
 
Tel (cell): 217 621-2137
 


From: fluen...@googlegroups.com <fluen...@googlegroups.com> on behalf of Sophie Gerday <fofi...@gmail.com>
Sent: Sunday, June 6, 2021 3:08 PM
To: FluencyBank <fluen...@googlegroups.com>
Subject: Help for translation
 
--

Yaruss, J Scott

unread,
Jun 6, 2021, 10:06:16 PM6/6/21
to Carol Hubbard Seery, Nan Bernstein Ratner, Sophie Gerday, FluencyBank
HI Sophie - Thanks for bringing this question to the group. It’s a really interesting one, and I appreciated seeing the responses about how others would do that coding. Due to the nature of my R01 (examining variability of observable stuttering behavior), I spend a lot of time these days thinking about how to code various observable speech behaviors, and it’s often a challenge.

One point really stood out to me in what Nan said: you’re not representing a diagnosis but rather just transcribing a behavior. So, you are observing an unfinished word or phonological fragment followed by a repetition of a complete word and then the completion of the unfinished word. That’s about all that we can really say about it.

What is interesting to me is this: We cannot know the *reason* for that unfinished word/phonological fragment. Perhaps the speaker discontinued that sound/word because they felt that they were about to stutter. In that case, what is really occurring here is stuttering, even though the observable behavior did not fit into the typical categories that are so often used for describing stuttering (or stuttering-like) behaviors. The internal state of the speaker simply isn’t available to us as a listener, so we can never know if the person was stuttering or not. We can only describe what we heard or saw on the surface. 

I think that’s why Nan and Carol both mentioned physical tension. Physical tension is a sign that the person may be struggling to get the word out, and that is a clearer marker that the person may have stuttered. Still, the absence of tension does not mean that the person did not stutter. We are learning more and more about the many behaviors and strategies that people who stutter use to mask or hide moments of stuttering so they are less observable to others. We simply have no way of knowing, based on the audio or even video sample, whether the speaker was doing that in this instance.

BUT this is why I’m writing, actually, because Carol mentioned something that seems to stand in contrast to what Nan said about not representing a diagnosis. Specifically, Carol, you mentioned that you “would not have classified the disfluent behavior with any category that might be considered stuttering-like/core/primary/within-word because there was no perceived tensionespecially as a non-stuttering speaker.” (Emphasis added by me.)

I’m curious: does this mean that if you were faced with a potentially ambiguous disfluency (as I believe that many are), you would have coded it as stuttered or not-stuttered based on the previously known diagnosis of the speaker as a person who stutters or not? Does that run the risk of introducing some confirmatory bias? (People who do not stutter do not exhibit stutter-like disfluencies very much because we wouldn’t have coded them that way knowing that they don’t stutter?) Perhaps I misread the comment. 

Also, is it possible for a person to exhibit a stuttering-like/core/primary/within-word disfluency that is actually a moment of stuttering even if there is no tension visible to the observer? (The answer to this is of course yes, so I’m wondering how those might have been captured or coded.) 

I know it’s been a while since those coding days for the project, and I don’t mean to put you on the spot. Really, my point in bringing this up is to explore the issue of how we handle ambiguous disfluencies (again, as many of them are because of the covert behaviors that people who stutter might engage in). I wonder about the consequences of calling them non-stuttered even when we cannot know for sure. (Likewise, I wonder about the common situation when a person experiences a sensation of being stuck under-the-surface but not show that in any way that a listener can perceive. We should be equally cautious about calling such things “fluent,” but that is a topic for a different post.)

That brings me back to Nan’s point of keeping it at the level of describing the behavior only and not of making a diagnosis — or even a judgment about the underlying nature of the behavior. (I think that’s one of the reasons that the terminology “stuttering-like” or “perhaps stuttered” or “possibly stuttered” is so important - that hedge gives the wiggle room that we need because we can never know for sure…)

Anyway, just some musings from someone who spends too much time thinking about these types of things… Thanks again for raising the issue and for the dialogue!

S




Vivian D. Sisskin

unread,
Jun 6, 2021, 11:20:27 PM6/6/21
to Yaruss, J Scott, Carol Hubbard Seery, Nan Bernstein Ratner, Sophie Gerday, FluencyBank

Thanks, Scott. Now the geeks (including me) will all come out of our bubbles :-)

I agree that it is important to code surface features at surface value. That said, I do code a whole word repetition differently based on qualitative features. If the whole word repetition is accompanied by tension or irregular timing, I code it as an SLD (stutter-like disfluency) and if it does not, I code it as an OD (other disfluency). However, this is not to say that coding it as an SLD would inform a diagnosis. There are too many other factors that go into diagnosis. Both those who stutter and those who do not may produce repetition that is tense. Think of the classic “Obama Stutter”. Clinically, we need to make inferences about what surface features represent, and in this case, it helps to generate multiple inferences for each observation. This is an exercise that all of my students do to help them think critically during assessment and diagnosis. We can infer that tension during repetition may point to a diagnosis of stuttering, but we can also infer time pressure, being interrupted, emphasis for meaning, emotional states, other conditions, etc. Likewise, phrase repetition (which I code as an OD) can point to language formulation, escape from stuttering, changing intended content, perfectionism (“saying it right”, as some of my autistic clients might say), holding the floor, repeating content with altered emphasis, etc.

Ultimately, why are we coding the sample? Most of us would agree that we can’t use this information alone to make judgments about diagnosis or therapeutic change. But, for the kind of therapy I do (for example), change can be inferred through shifts in disfluency types, so coding them with clear rationale is helpful. 


Hope everyone is doing well. I miss the in-person interaction!


Vivian


Vivian Sisskin, Clinical Professor
Dept. Hearing and Speech Sciences
University of Maryland
0100 LeFrak Hall
College Park, MD 20742

ASHA Fellow
Board Certified Specialist in Fluency Disorders

Sisskin Stuttering Center


Yairi, Ehud

unread,
Jun 7, 2021, 12:21:29 AM6/7/21
to Yaruss, J Scott, Carol Hubbard Seery, Nan Bernstein Ratner, Sophie Gerday, FluencyBank

Well Ms. Gerday,


At this point in time, you should sleep well in spite of this tumul concerning a rather small issue at this point of time. I can tell you that thousands of highly sophisticated MRI tests end up with questionable diagnosis or misdiagnosis. So, we have our share.

As for Scott’s reply, it is a bit on the long side (may have inspired Shakespear to write an additional act to “Much to do about nothing.”) But, he makes a point or two to consider.


In the meantime, may I refer you to page 100 in the Yairi and Ambrose “Early Childhood Stuttering” (2005):   

 

 As can be seen, the various proposed global schemes are quite similar, reflecting a general agreement in the field about the nature of the different types of disfluency. Still, they seem to convey the realization that a distinction between what is stuttering and what is normal is not always possible. We are aware of these limitations and their implications for the data that they generate. Practically, one may assume that some errors occur in both directions: counting as stuttering some disfluencies that actually were normal, and counting as normal some disfluencies that were stuttering. We agree with Conture (2001) that, for the time being, it might be necessary to recognize and accept a relatively small percentage of error. As the state of knowledge about stuttering increases, and with the development of more sensitive, automated technology (Howell et al., 1997), it should become possible to derive more accurate measures.” 

 

So dear, as I said, sleep well. At age 82, I am rather relaxed about all this.


Ehud

__________________


Ehud Yairi
Professor Emeritus
University of Illinois, College of Applied Health Sciences, Speech & Hearing
Tel Aviv University, Faculty of Medicine, Communication Disorders
 
Tel (cell): 217 621-2137
 


From: fluen...@googlegroups.com <fluen...@googlegroups.com> on behalf of Yaruss, J Scott <j...@msu.edu>
Sent: Sunday, June 6, 2021 9:06 PM
To: Carol Hubbard Seery <cse...@uwm.edu>
Cc: Nan Bernstein Ratner <nra...@umd.edu>; Sophie Gerday <fofi...@gmail.com>; FluencyBank <fluen...@googlegroups.com>

Howell, Peter

unread,
Jun 7, 2021, 3:18:22 AM6/7/21
to Sophie Gerday, FluencyBank

What we do is pre-process speech using our version of Levelt’s interruption/repair parser. That would remove events like this repair (I agree with Ehud about characterization of this event). I think that the other main benefit of doing the parsing is that it deals with phrase repetitions in a similar way (excludes them as stutters). Indeed, we also parse to remove whole-word repetitions (indication of covert repairs according to Levelt).

Peter   

--

You received this message because you are subscribed to the Google Groups "FluencyBank" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fluencybank...@googlegroups.com.

Sophie Gerday

unread,
Jun 7, 2021, 8:56:25 AM6/7/21
to FluencyBank

Good morning all,

Thank you all for your answers and your reflective developments. It is very rich in exchange! As Mr Yaruss says very well, this coding is a real challenge!

I slept well, thank you Mr. Yairi!

Anyway, reflection remains in order and the final decision will be taken after discussion ...

Thank you all for your quick reactions.

Take care of yourself.


Sophie

Rosalee Shenker

unread,
Jun 7, 2021, 9:23:41 AM6/7/21
to Sophie Gerday, FluencyBank
I just read the thread so I am coming a bit late to the party.  I appreciated reading everyone's thoughtful responses.  I am not a geek, but I am an 'old person' and as a member of that club Ehud's response really resonated with me.  Thank you all for inspiring me on this hot Monday morning in Montreal.

Rosalee

Rosalee C. Shenker, Ph.D., CCC-SLP
Founder
Montreal Fluency Centre
4626, rue Sainte-Catherine Ouest, Westmount, Qc H3Z 1S3 
Tél. 514.489.4320 poste 242 
rosalee...@montrealfluency.com
www.montrealfluency.com
www.facebook.com/MontrealFluency

Centre d’orthophonie et de psycho-éducation bilingue et à but non-lucratif offrant à la communauté des services d’une qualité exceptionnelle depuis plus de 25 ans. Des bourses sont disponibles pour les familles dans le besoin. 
/
Not-for-profit, bilingual centre offering exceptional quality speech-therapy and educational services to the community for over 25 years; bursaries are available to families in need.  

Faire un don / Give now 
BN: 884489766RR0001


Kenneth St Louis

unread,
Jun 7, 2021, 9:45:40 AM6/7/21
to Vivian D. Sisskin, Yaruss, J Scott, Carol Hubbard Seery, Nan Bernstein Ratner, Sophie Gerday, FluencyBank

Yep, Scott and Ehud laid it out pretty well. It is impossible to develop an a priori coding system to know with absolute certainty whether phrases repetitions are stuttered or not. I remember a guy I ran in my dissertation who had no observable physical tension or evidence of tension in his voice or sound production except perhaps a short pause before starting over. He would say things like “We went to the- we went to the- we went to the- we went to the store.” I counted these as stutters and I am 100% certain that he stuttered in these situations. Maybe it was the number of phrase repetitions that convinced me, but I probably could not code that for certain either.

 

Ken

 

 

****************************************

Kenneth O. St. Louis, Ph.D., BCS-F

Professor Emeritus

Dept. Communication Sciences & Disorders

805 Allen Hall, PO Box 6122

West Virginia University

Morgantown, WV 26506-6122

Logan,Kenneth J

unread,
Jun 7, 2021, 9:59:35 AM6/7/21
to Rosalee Shenker, Sophie Gerday, FluencyBank

HI all,

 

I am even later to the party. I didn’t read every word of the exchange below, but my take on it would be similar in some ways to what some folks have said, and perhaps different from others.

 

Disfluencies like the one you point out below highlight the limitations of convention disfluency classification systems – in short, there are not enough labels to capture all that one encounters when transcribing speech!

 

In my view, this type of disfluency has elements of both phrase and part-word repetition and thus there is no consensus agreement on how to describe its structure. The moment of interruption in speech occurs between the onset and rime of a syllable. Thus, in Ed Conture’s taxonomy it has elements of “within-word” disfluency, which makes it like part-word repetition and sound prolongations, and thus “stutter-like.” It differs from a classic part-word repetition in that, instead of starting the repair at the beginning of the syllable [as is the case in sound repetition], the speaker retraces back a bit more, to a previous syllable.

 

I think Nan made this comment right off the bat, but what I’ve done in my research that compares children who stutter with typical children is simply to count how often this type of disfluency occurs (at a group level, it’s not very often, for either type of speaker) and then run statistical comparisons to see if children who stutter produce this type of disfluency more often than children who do not stutter (they do!!).

  • I don’t think that this narrow finding ever made it into any of my published studies, but we did look at it in preliminary data analysis in a couple of studies, and on the basis of those analysis, decided to classify into our “stutter-like” bin. The children who stutter produce more of them.
  • That is not the same as saying that every one of these was perceived as an instance of stuttering in either group. It’s just that disfluencies with this type of structure were produced significantly more often by the children who stutter than by the typical children, and in that narrow sense are “stutter-like.” Based on the work from Ehud’s lab, I would also expect that children who stutter would to produce more multi-iteration instances of these than typical children.
  • For the reasons above, we opted to include this type of disfluency under scorable repetition types in the Stuttering Severity Instrument-5 (SSI-5), which Pro-Ed currently has out for norming.

 

As others on this email chain have noted, in the case of an individual speaker, however, the occurrence of an instance or two of this type of disfluency doesn’t really tell a clinician much. It needs to be examined in the broader context of everything else the speaker does (e.g., total disfluencies, disfluency duration, iterations per repetition, etc. etc, etc.)

 

Hope this is useful.

 

Best,

 

Ken

 

 

Kenneth J. Logan, Ph.D., CCC-SLP

Associate Professor, Interim Chair

Department of Speech, Language, and Hearing Sciences

1225 Center Drive, Room 2147A

University of Florida

Gainesville, FL 32610-0174

 

Tele. 352-273-6561

 

 

 

 

From: <fluen...@googlegroups.com> on behalf of Rosalee Shenker <rosalee...@gmail.com>


Date: Monday, June 7, 2021 at 9:24 AM
To: Sophie Gerday <fofi...@gmail.com>
Cc: FluencyBank <fluen...@googlegroups.com>

Subject: Re: {SPAM?} Help for translation

 

[External Email]

Luc De Nil

unread,
Jun 7, 2021, 10:18:11 AM6/7/21
to Kenneth St Louis, Vivian D. Sisskin, Yaruss, J Scott, Carol Hubbard Seery, Nan Bernstein Ratner, Sophie Gerday, FluencyBank

Hi all, interesting discussion.

 

I agree completely with Nan and others that we have to keep in mind that we are transcribing observable behaviour and need to keep our assumptions to a minimum. After all, a diagnosis or ‘stuttering or not’ is (needs to be) based on much more than a pure disfluency count. I also agree wholeheartedly with Ehud that we have to accept that our count will not be 100% accurate and that errors of misinterpretation will be made if we try to establish a disfluency frequency baseline. We need to keep in mind that for instance on a 500 word sample, the difference between 12% stuttering and 10% stuttering is 10 stutters, and I would argue that the difference between 12% and 10% is not clinically significant. So if we disagree on the classification of 10 instance of disfluency, it may not be a big deal (hence Ehud’s recommendation to have a good sleep 😊)

 

In this particular example. If the child says “he’s m… he’s making noise”, we cannot automatically assume that the child is trying to say “he’s m(aking)…). What if the child were planning to say “he’s m(oving)… he’s making noise”, so transcribing it as m(aking) is making an assumption that may be correct but also may be wrong. So, better to stick with observable behaviour.

 

Also, can we please stop using SLD? If we say that something is ‘stuttering-like’ this means that it sounds like ‘stuttering’, which of course means that there is something else that we call ‘true’ stuttering but this instance is only ‘stuttering-like’. An analogy is ‘meat’ and ‘meat-like’. The latter may look and taste (a bit) like meat but it is not real meat. To bring it a bit closer to home, if clinicians talk about ‘Parkinson-like symptoms’, they refer to symptoms that look like PD but are not the result of PD. I find the use of SLD utterly confusing and always find it difficult to explain to my students. If we want to indicate instances of disfluency that may be stutters but we are not completely sure (as is the case in this example), maybe we can call it ‘stuttering-likely” but I would not call it ‘stuttering-like’.

 

Luc

 

__________________________________________

Luc De Nil, PhD

Professor

Department of Speech-Language Pathology

University of Toronto

luc....@utoronto.ca

Tel. 416-978-1789

 

If you have an accommodation need for a scheduled meeting,

please email me directly and I will do my best to make appropriate arrangements.

 

 

 

 

From: fluen...@googlegroups.com <fluen...@googlegroups.com> on behalf of Kenneth St Louis <Ken.S...@mail.wvu.edu>
Date: Monday, June 7, 2021 at 9:45 AM
To: Vivian D. Sisskin <vsis...@umd.edu>, Yaruss, J Scott <j...@msu.edu>
Cc: Carol Hubbard Seery <cse...@uwm.edu>, Nan Bernstein Ratner <nra...@umd.edu>, Sophie Gerday <fofi...@gmail.com>, FluencyBank <fluen...@googlegroups.com>
Subject: Re: Help for translation

EXTERNAL EMAIL:

Mark Irwin

unread,
Jun 8, 2021, 10:27:21 AM6/8/21
to Logan,Kenneth J, Rosalee Shenker, Sophie Gerday, FluencyBank
Hate to miss a party.....especially one with such an impressive guest list as this.

Very interesting and thoughtful responses with which I am sure everyone agrees. 

A slight diversion if I may......I hope also we can all agree that whatever skills we have in data collection, science works best when data collection is aimed at testing an hypothesis. As Bill Perkins said in his 1995 book Stuttering and Science...
 " When our inventive and strategic thinking becomes equally as strong as our empirical-testing ability, then increased progress in our understanding of stuttering can be expected" (p25).  Perkins of course disliked seeing the continuing focus on data collection for its own sake.....he saw it as an inefficient fishing (for an hypothesis) expedition.

If anyone is interested in contributing to inventive and strategic thinking on this subject then please send me an email.
Are we able to work together to make the future better than the past? "To be or not to be?" 

Mark Irwin






M.C.J.P. Franken

unread,
Jun 14, 2021, 9:03:51 AM6/14/21
to Luc De Nil, Kenneth St Louis, Vivian D. Sisskin, Yaruss, J Scott, Carol Hubbard Seery, Nan Bernstein Ratner, Sophie Gerday, FluencyBank

Hi everyone,

I agree, very interesting discussion, and I also agree with Luc, and others about what has been said about the importance of keeping our assumptions to a minimum when we are describing what we perceive. Ehud mentions Conture’s volume, but more recently, Clark et al 2013 have shown that we need to accept there are borderline cases, and that there are cases that we cannot classify based on perceptual data only (reference below).

 

Clark CE, Conture EG, Walden TA, Lambert WE. Speech sound articulation abilities of preschool-age children who stutter. J Fluency Disord. 2013 Dec;38(4):325-41. doi: 10.1016/j.jfludis.2013.09.004. Epub 2013 Oct 9. PMID: 24331241; PMCID: PMC3868004.

 

Besides, I agree with Luc:  I am not happy with the term stuttering like dysfluencies either. However, I am not sure whether I agree with the statement by Ken that “the moment of interruption in speech occurs between the onset and rime of a syllable”.  Did you investigate this ? I would be very interested to read about this. I can see that your statement may be true for sound repetitions and prolongations, but how about an instant block on the very first sound of a word, without realizing the very first sound ?  And how about a block on a word initial vowel?

 

Ken, feel free to respond only to me, because this may discussion is outside the scope of the ongoing discussion.

 

Best,

Marie-Christine

Logan,Kenneth J

unread,
Jun 14, 2021, 10:30:57 AM6/14/21
to M.C.J.P. Franken, Luc De Nil, Kenneth St Louis, Vivian D. Sisskin, Yaruss, J Scott, Carol Hubbard Seery, Nan Bernstein Ratner, Sophie Gerday, FluencyBank

Hi Marie-Christine,

 

Thank you for your comments.

 

To clarify, my comments in the previous email were in reference to disfluencies that have a structure like this: …[the b-]  the boy went… . In such disfluencies, speech is interrupted in mid-syllable and the subsequent repair features a retrace to the beginning of “the”.

 

In exploratory data analysis for some of the studies I conducted in the mid-1990s to early 2000’s, we found that speakers who stuttered produced more of this type of disfluency than the speakers with typical fluency. I would need to look back at those articles, but as I recall, in that era, we (like many other people) were making the “stuttering-like” versus “non-stuttering like” distinction in studies that compared fluency characteristics of talker groups, and for that reason, we needed to know if that type of disfluency was something that people who stutter did more often than typical speakers.

 

In subsequent years, however, I have moved away from making the stuttering-like vs. non-stuttering-like distinction. As others on this email chain seem to do, I have gone toward simply describing the types of disfluency that are produced. We used a descriptive approach like that, for example, in the Test of Childhood Stuttering. As others have noted in this email chain, this approach makes no assumptions about whether specific instances of disfluency might be judged as “stuttering” or not.

 

I hope this clears things up.

 

Best,

 

Ken

 

From: <fluen...@googlegroups.com> on behalf of "M.C.J.P. Franken" <m.fr...@erasmusmc.nl>
Date: Monday, June 14, 2021 at 9:04 AM
To: "Luc....@utoronto.ca" <luc....@utoronto.ca>, Ken StLouis <Ken.S...@mail.wvu.edu>, Vivian Sisskin <vsis...@umd.edu>, J Yaruss <j...@msu.edu>
Cc: Carol Hubbard Seery <cse...@uwm.edu>, Nan Ratner <nra...@umd.edu>, Sophie Gerday <fofi...@gmail.com>, FluencyBank <fluen...@googlegroups.com>
Subject: RE: Help for translation

 

[External Email]

Hi everyone,

Nan Bernstein Ratner

unread,
Jun 14, 2021, 11:43:40 AM6/14/21
to Logan,Kenneth J, M.C.J.P. Franken, Luc De Nil, Kenneth St Louis, Vivian D. Sisskin, Yaruss, J Scott, Carol Hubbard Seery, Sophie Gerday, FluencyBank
Devoid of other features, the retrace that Ken uses as an example is a very good example of Levelt's principle that when constituents are interrupted for whatever reason, the speaker tends to reinitiate the entire planning unit.

N

Nan Bernstein Ratner, F-, H-ASHA, F-AAAS, ABCLD
Professor
Hearing and Speech Sciences
University of Maryland
0100 Lefrak Hall
College Park, MD 20742


President, International Fluency Association (IFA; http://theifa.org)
Director, University of Maryland Autism Research Consortium (UMARC), www.autism.umd.edu
Faculty, Language Science (languagescience.umd.edu; Neuroscience & Cognitive Neuroscience (NACS, nacs.umd.edu), Developmental Science Field Committee



Lasalle, Lisa

unread,
Jun 14, 2021, 1:10:08 PM6/14/21
to Nan Bernstein Ratner, Logan,Kenneth J, M.C.J.P. Franken, Luc De Nil, Kenneth St Louis, Vivian D. Sisskin, Yaruss, J Scott, Carol Hubbard Seery, Sophie Gerday, FluencyBank
Good exchange to catch up on.  In LaSalle & Conture 1995, we found that 30 children who stutter produced significantly more phrase repetitions with cutoff words then did normally fluent peers — as in  …[the b-]  the boy went… .
This was from my dissertation and the Covert Repair Hypothesis served as the theoretical framework. 

Disfluency clusters of children who stutter: Relation of stutterings to self-repairs

Lisa R LaSalle, Edward G Conture
Journal of Speech, Language, and Hearing Research 38 (5), 965-977, 1995
Hope that helps some with Ken’s and Nan’s comments,

Lisa


Lisa LaSalle, Ph.D., CCC-SLP
Professor and Chair
Communication Sciences & Disorders Department
University of Redlands
1200 Colton Ave
Redlands, CA 92373
(909) 748-8163
---------------------


On Jun 14, 2021, at 8:45 AM, Nan Bernstein Ratner <nra...@umd.edu> wrote:



EXTERNAL EMAIL: Proceed with caution when clicking on links or opening attachments.

M.C.J.P. Franken

unread,
Jun 14, 2021, 1:20:44 PM6/14/21
to Logan,Kenneth J, Luc De Nil, Kenneth St Louis, Vivian D. Sisskin, Yaruss, J Scott, Carol Hubbard Seery, Nan Bernstein Ratner, Sophie Gerday, FluencyBank

Thanks Ken!

Yes, it certainly helped that you clarified that your comment was in reference to the specific type of dysfluency “…[the b-]  the boy went…”, to describe this disfluency more precisely. I had interpreted it as a general statement about moments of stuttering.

 

Best,

Marie-Christine

M.C.J.P. Franken

unread,
Jun 14, 2021, 1:43:52 PM6/14/21
to Lasalle, Lisa, Nan Bernstein Ratner, Logan,Kenneth J, Luc De Nil, Kenneth St Louis, Vivian D. Sisskin, Yaruss, J Scott, Carol Hubbard Seery, Sophie Gerday, FluencyBank

Thanks for bringing your study to our attention Lisa.

I wish we were able to compare the speech motor characteristics of the phrase repetitions in your study 😉

 

Best,

Marie-Christine

Jaeger, Jeri

unread,
Jun 14, 2021, 5:03:11 PM6/14/21
to Nan Bernstein Ratner, Logan,Kenneth J, M.C.J.P. Franken, Luc De Nil, Kenneth St Louis, Vivian D. Sisskin, Yaruss, J Scott, Carol Hubbard Seery, Sophie Gerday, FluencyBank

​Please see my book Kids' Slips for a thorough discussion of this issue (young children's slips of the tongue).


From: fluen...@googlegroups.com <fluen...@googlegroups.com> on behalf of Nan Bernstein Ratner <nra...@umd.edu>
Sent: Monday, June 14, 2021 11:44 AM
To: Logan,Kenneth J
Cc: M.C.J.P. Franken; Luc De Nil; Kenneth St Louis; Vivian D. Sisskin; Yaruss, J Scott; Carol Hubbard Seery; Sophie Gerday; FluencyBank

Martine Vanryckeghem

unread,
Jun 18, 2021, 5:53:03 AM6/18/21
to Luc De Nil, Kenneth St Louis, Vivian D. Sisskin, Yaruss, J Scott, Carol Hubbard Seery, Nan Bernstein Ratner, Sophie Gerday, FluencyBank

Dear colleagues,

As a former colleague would say “here are my 2 c”.

I want to expand on two things Luc indicated.

We try to diagnose if someone is a person who stutters or not. We all know that the stutterER is more than stutterING, which I have often emphasized in my publications. Indeed, a count of dysfluencies is but one component of what occurs within the individual who stutters. And, we all know too well that the way stuttering has an impact on the person depends on much more than just the dysfluency count.

 

I also agree with Luc as it relates to SLD’s. I am so happy you brought this up, Luc. It is a term that I discuss with my students and I personally do not use. I ask my students to carefully look at the “SLD” definition and to indicate what the difference is with the behaviors that are typically defined as “stuttering”. It is also always an examination question, and even after discussion, most students still think that both terms mean different things. I have had several discussions over the past years with colleagues about the term which, in my opinion (like Luc’s) is a misnomer. When someone talks about “…-like” (e.g. autistic-like), it means that someone has tendencies for a particular disorder, but “it” does not exactly meet the criteria for that disorder. On a continuum, it is approaching it, has some characteristics, but does not correspond to the actual disorder. It is more like having “tendencies” toward the disorder rather than the disorder itself.

 

Best wishes,

Martine

Reply all
Reply to author
Forward
0 new messages