Two part question:

First, I want to find out if people in North America (the U.S. and
Canada, either/or, preferably both) are speaking (and therefore
listening) faster now than they were 20 to 40 years ago. I need this
documented from a credible source, as I intend to use it in a
manuscript. A study by linguistics researchers would be ideal.
However, since most people's daily rate-of-speech is rarely documented
or counted, an acceptable proxy might be if someone has counted the
rate of speech of newscasters on television shows now and 20-40 years
ago. Scripted shows, such as sitcoms, wouldn't count as I understand
that producers have recently started increasing the speed of dialog in
an attempt to make their characters seem more intelligent and their
plots seem "edgier." This, therefore, would not represent the speech
patterns of normal people.

Part two: Is there any credible research indicating that there are
"hardware" limits on how fast (in words per minute) people can hear
speech and still comprehend it? Is there a normal distribution of
abilities in this area (i.e., a few people can understand very rapid
speech, most can understand speech of "X" words a minute, and a few
can only understand very slow speech)?

What I'm trying to ascertain with these questions relates to
information and sensory overload. Everyone complains about overload -
is it really happening? Rate of speech is one aspect of this. There
may be
additional questions, depending on the results of this search.

Hi futurist-ga,

This is a fascinating research topic.

It seems that the studies on speed do not acknowledge any across the
board speeding up of the normal speech patterns in the US and Canada.
Researchers have been studying the ability of people to process
time-compressed speech for decades and there has been no notice of a
higher baseline.

Some effects of speaking rate on phonetic perception.
This is one of the seminal papers on speaking rate published in 1981
by JL Miller (jlmiller@neu.edu). His earliest papers on linguists were
published in the 1950s.
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=7267718&dopt=Abstract

Internal structure of phonetic categories: effects of speaking rate.
Miller, who is still focused on effects of speaking rate on
perception, published this study in 1997. He does not indicate
awareness of any change in the “normal” speaking rate.
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=9396164&dopt=Abstract

A Timing Model for Fast French 
While this article does not address changes in the rates of speech
specifically, the references go back into the 1970s and there does not
seem to be any indication that there has been an overall change. The
group studies “Models of speech timing.” It would seem that if there
had been an increase in the speed of normal speech, they would have
noticed. The only problem is that this study is based on spoken French
– not English.
http://cogprints.ecs.soton.ac.uk/archive/00000885/00/KellerZellnerTimingModel.pdf
http://cogprints.ecs.soton.ac.uk/archive/00000894/00/ZellnerTemporalStr.pdf
is another article by the same group.
###

The problem with documenting newscasters speaking rates would fall
into the same category as sit-coms. The management dictates the styles
and the actors (broadcasters) provide whatever the ratings gurus
recommend.

US sitcom stars told to speak faster
This discusses the sit-com speed speech phenomenon.
http://www.ananova.com/entertainment/story/sm_710061.html

Motivations for News Language Style: Audience Perception or Cultural
Orientation
This article does not specifically address the “rate” of speech in
news media. However, the imposition of “styles” on the broadcasters
seems to put them in the same category as sit-com actors. If they are
not using normal speech patterns on air, they can’t be used to
determine if normal speech patterns have shifted.
http://www.criticism.com/md/newslang.html
###

The following articles address the cognitive processes of language
recognition. There are many parts to processing words and sentences.
Visual cues such as gestures and facial movements account for much of
our understanding of the spoken word. We assign themes and are primed
by context and experience to fill in what we do not actually hear. One
study found that even two year olds can identify a word from hearing
only the first 300ms.

When half a word is enough: infants can recognize spoken words using
partial phonetic information.
Infants as young as two can recognize words after hearing only the
first 300ms.
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11480931&dopt=Abstract

Pauses and ryhthm are helpful in comprehension, but studies using
automated speech with these removed prove that they are not necessary.

Techniques, Perception, and Applications of Time-Compressed Speech 
“The normal English speaking rate is between 130-200 words per minute
(wpm). When speaking fast, a talker unintentionally changes relative
attributes of his speech such as pause durations, consonant-vowel
duration, etc. Talkers can only compress their speech to about 70%
because of physiological limitations.”
http://xenia.media.mit.edu/~barons/avios92.html

Using the bell curve, the statistical representation which applies to
most human attributes, the 130-200 range would apply to 90% of the
population. There would be 5% speaking more slowly or not at all - due
to genetic or other factors and 5% speaking faster.

There are many additional studies on the physiological and cognitive
aspects of speech processing.

Speech Insights Sound Off in the Brain
"Brain cells react in a unique pattern and with lightning speed to
so-called prosodic features of speech, such as drawing out a word,
pausing for emphasis, and raising the pitch of one's voice, reports a
team of neuroscientists headed by Karsten Steinhauer of the Max Planck
Institute of Cognitive Neuroscience in Leipzig, Germany. Immediate
neural responses to these speech modulations allow listeners to get a
head start in decoding the sometimes fuzzy meanings of sentences,
Steinhauer and her colleagues contend."
http://www.findarticles.com/cf_0/m1200/5_155/54031951/p1/article.jhtml

Speech loses beat in dyslexia. 
Reports on a study that found a "sound cue, which lasts for one-tenth
to one-fifth of a second, marks the transition from a consonant sound
to a speech segment beginning with a vowel." Missing this brief
transition interferes with word recognition and is symptomatic of
dyslexia.
http://www.findarticles.com/cf_0/m1200/6_162/90870810/p1/article.jhtml

Speech insights spark statistical static.
“However, when the eight nonsense words were separated with
25-millisecond gaps, a third group proved able to generalize about
word structure and identify the ninth nonsense word as wordlike.
Although participants reported no awareness of these fleeting pauses,
such cues made the speech stream more similar to the rhythm and
intonations of natural speech, Mehler asserts.”
http://www.findarticles.com/cf_0/m1200/9_162/91915105/p1/article.jhtml

Prosodic boundaries, comma rules, and brain responses: the closure
positive shift in ERPs as a universal marker for prosodic phrasing in
listeners and readers.
Spacing of words and segmentation of sentences influences the
comprehension of both readers and listeners.
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11523275&dopt=Abstract

Building Blocks of Talk
"From a dynamic-systems perspective, infants and young children learn
to communicate by seamlessly combining two movement systems, one for
speech and the other for gestures, according to Jana M. Iverson of the
University of Missouri in Columbia and Esther Thelen of Indiana
University at Bloomington. It's time to reconsider the popular notion
of gestures as a decorative sideline to the real business of saying
what you mean, the two psychologists argue. Instead, gabbing and
gesticulating go hand in hand."
http://www.findarticles.com/cf_0/m1200/22_157/62981021/p4/article.jhtml

Eye movements and lexical access in spoken-language comprehension:
evaluating a linking hypothesis between fixations and linguistic
processing.
More data on the use of visual cues in language processing.
“Experiments using this paradigm have shown that visually presented
referential information, including properties of referents relevant to
specific actions, influences even the earliest moments of syntactic
processing.”
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11196063&dopt=Abstract

Native signers may get right-brain aid.
“Use and understanding of spoken language is largely orchestrated by
the brain's left side. However, a team of neuroscientists now suggests
that a right-brain area assumes a critical role in deciphering sign
language, at least among native signers.”
“Brain-damaged patients provide contrasting clues to the biology of
language, Hickok notes. For instance, right-brain injuries usually
spare language skills in fluent signers. But some studies suggest that
damage to the right angular gyrus and nearby tissue weakens the grasp
of complex sentences presented either in writing to hearing patients
or as signs to deaf ones.”
http://www.findarticles.com/cf_0/m1200/24_160/81827795/p1/article.jhtml

Making sense during conversation: an fMRI study.
A study discovered activity in both brain hemispheres when “themes”
and words were processed.
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11733724&dopt=Abstract

Syntactic priming in spoken production: linguistic and temporal
interference.
Priming is an important effect in language processing. This effect is
more pronounced and longer lasting in spoken language.
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11219957&dopt=Abstract

Contribution of speech rate to speech perception in multichannel
cochlear implant users.
“The results reveal that the rate of speech is an important factor in
improving the speech perception of cochlear implant users.”
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=12184594&dopt=Abstract

Linguistic and nonlinguistic auditory processing of rapid vowel
formant (F2) modulations in university students with and without
developmental dyslexia.
Although the purpose of this study was to determine the ability of a
dyslexic population to process language, the findings in the control
group are of interest to this topic. Normal controls improved
performance with repetition indicating that speed was not a distractor
in normal language processing.
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=12030500&dopt=Abstract

The relationship between the intelligibility of time-compressed speech
and speech in noise in young and elderly listeners.
It seems that speeding up speech in a noisy background makes language
perception more difficult for older subjects, but not for younger
ones.
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11831813&dopt=Abstract

Effects of time pressure on mechanisms of speech production and
self-monitoring.
This study did not find that speeding up speech interfered with
processing.
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11385824&dopt=Abstract

Speaking speed effects on delayed auditory feedback disruption of
speech fluency.
“From these results, we suggest that, when speaking more quickly, one
uses more central mechanisms of movement programming
(cortical-cerebellum-thalamus-cortical, cortical-corpus
striatum-thalamus-cortical, and cortical-thalamus-cortical circuits),
or attentional control (cortico-reticular-cortical circuits) than
peripheral mechanisms (tactile, proprioceptive, and acoustic
circuits).”
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=10710757&dopt=Abstract

Speed of processing in normal aging: effects of speech rate,
linguistic structure, and processing time.
Older adults showed decreased comprehension, “significantly steeper
rates of performance decline with increasing speech rate.”
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=4031406&dopt=Abstract

Effect of word speed, number of words, and habituation on the
discrimination score of synthesized speech.
In synthesized speech, the speed did not have any effect on the
processing. Could this be due to the absence of visual cues making
synthesized speech at any speed as difficult to process?
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=6444503&dopt=Abstract

Perceptual Adjustment to Time-Compressed Speech: A cross-linguistic
study
This study looks at the ability to process time-compressed speech in
several languages. The interesting thing about the compresses speech
is that after 10 to 15 repetitions, the comprehension is good and
remains at that level. This would seem to support an argument that
faster speaking rates can be internalized without interfering with
comprehension.
http://cogprints.ecs.soton.ac.uk/archive/00000930/00/comp97.pdf

Auditory Analysis in Spoken Language Perception
Loudness, rather than speed seemed to be the main variable influencing
perception in this study.
http://www.hip.atr.co.jp/departments/Dept1/progress94/node7.html

One last reference that seemed interesting...

Traditional linguists is modeled on the "classic user" who is assumed
to have certain characteristics. This paper discusses whether the
"real" user bears any resemblance to the model.

Real Language Users
"Psycholinguistic research has so far explored one of the
possibilities suggested by Chomsky and Miller - the notion of an
infinitely competent but capacity limited language user. There is
another possibility still to be explored - that of a finitely
competent language user with an infinite capacity to learn."
http://cogprints.ecs.soton.ac.uk/archive/00000712/01/real.html

Perhaps it will fuel a new line of research that will find that we can
expand our capabilities much more than traditional research expects.

In terms of the current research, if there is an overload coming from
speeded up speech, it's not obvious to academia. The artificial world
of the media may be introducing quicker speech. The world of
artificial intelligence and data communications is definitely trying
to speed us up. However, it doesn't seem to have taken, yet...

Search terms: 
Speech perception rate spoken word recognition
rate speed speech processing

Great question!

bcguide-ga

---

Request for Answer Clarification by futurist-ga on 26 Nov 2002 09:15 PST

Thanks for the citations. Almost all of them relate to the structure
of language, and classic linguistic studies on how speed, duration,
pauses, and spacing affect comprehension in both human and synthesized
speech, which is peripherally related to my question, but doesn't
actually answer it. Or perhaps I've missed something (which is always
possible).

So let me ask simply whether I've missed the answers I was looking for
in the material you sent:
1) Did you find any evidence that people are speaking faster today
than they did 20-40 years ago (not counting scripted television
shows)?
2) Did you find any evidence that there are specific limitations in
understanding more rapid speech?

On the second point, you seem to have found two relevant data points:
older people have a harder time comprehending faster speech than
younger people. And most linguists seem to consider speech patterns of
130-200 wpm to be roughly the norm, from which it is tempting to infer
(but which is not demonstrated) that this is the "natural"
comprehension rate.

Look forward to hearing.

Thanks.

---

Request for Answer Clarification by futurist-ga on 29 Nov 2002 07:59 PST

I sent an earlier request for clarification on Tuesday. I know this is
a short week, but I am working on a manuscript, and I do have a
deadline. Perhaps you can let me know if you expect to have further
clarification, as per my earlier request for clarification, and if so,
roughly when.

Thanks.

---

Clarification of Answer by bcguide-ga on 30 Nov 2002 14:31 PST

Hi,

Sorry for the delay in responding due to the holiday here in the US. 

Searching online and off has produced no research that indicates
normal speech patterns are speeding up. There have been volumes of
research on speaking, listening and comprehension speeds in the fields
of neuro-cognition, linguistics, human factors and synthetic
voice/data processing. With all of the studies, no one has noted any
change in the base rate of normal speech. Although there is no study
that specifically looks at comparisons of the speed of speech over the
last 40 years, one would expect that the phenomenon would have been
noted if it were present. If people were speaking faster the baseline
would have had to be raised and there is no indication that this is
the case. Studies from the 50s and the most recent studies all use the
same figures for "normal" human speech.

As indicated in the previous citations, the "average" rate is 130 -
200 wpm. Some citations that relate to the "natural comprehension
rate" are:

Compare that to the average speed of speech, at up to 185 words per
minute.
http://www.hamilton.net/relay/wi/fastran/

People comfortably can hear words that are spoken at from 150 to 160
words per minute (Williams, 1998).
http://www.humanfactors.com/downloads/aug00.asp

Abstracts of papers that discuss physiological limitations on the
rates of human speech production (not comprehension):
Digital, labial, and velopharyngeal reaction times in normal speakers.
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=3168264&dopt=Abstract

Control of rate and duration of speech movements.
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=3882804&dopt=Abstract

This one is a neural network model of:
Speech Sound Acquisition, Coarticulation, and Rate Effects in a
Neural Network Model of Speech Production
http://cns-web.bu.edu/~guenther/psychrev95.pdf
which provides a cognitive basis for limitations on the production of
speech sounds - although no figures for the limits are proposed.

No such constraints exist for speech that is time compressed using
computer models so the limits of speech production in humans do not
have to dictate speech comprehension...

http://citeseer.nj.nec.com/cache/papers/cs/1731/http:zSzzSzwww.ehess.frzSzlscpzSzpersonszSzdupouxzSz_pucompr.pdf/dupoux96perceptual.pdf
The PDF file is a poor copy, but it is readable. The study discusses
comprehension of speech that was time compressed up to 38% that was
able to be understood and retained to some degree after multiple
repetitions.

On the Role of Space and Time in Auditory Processing
http://www.isr.umd.edu/Labs/NSL/News/Shamma_TCS_Aug.pdf

discusses some of the current theory in the field.

There is no clear understanding of how the brain manages to transform
language signals into data that can be stored. Until we know how the
system works, it is difficult to put a limit on the speed at which it
can work.

So, in answer to your clarification request...

So let me ask simply whether I've missed the answers I was looking for
in the material you sent:
1) Did you find any evidence that people are speaking faster today
than they did 20-40 years ago (not counting scripted television
shows)? No, there seems to be no evidence to back up this theory.

2) Did you find any evidence that there are specific limitations in
understanding more rapid speech?
No, the models of auditory processing are not clear enough to test the
absolute limits. It seems that given enough repetition, the brain
systems are plastic enough to compensate for faster input or partial
input.

It may not be the data you were hoping to find, but that's the answer.

If you need further clarification, please post. I will get back to you
right away...

Regards,
bcguide-ga

The researcher's answer was comprehensive and complete. It disproved,
or at least did not support, my hypothesis, and therefore was a
disappointment to me. That's not his/her fault; negative results are
an important part of any honest research, even if I was hoping for
something different. I appreciate his/her honesty.

As a policy debate coach, I have observed that humans can understand
speech even at extremely rapid rates.

Some types of academic debate have evolved to include the practice of
rapid speech and speed reading, as debaters attempt to increase the
quantity of arguments in the debate.

It has gotten to the point where a lay-person would have a very
difficult time understanding most of the debate, but people
experienced with the activity seem to have little difficulty.

Thus, it seems that we can adapt and learn to understand rapid speaking,