|
about faq's links products |
 |
Welcome
to Semantic Compaction Systems,
*Registered Trademark of Semantic Compaction Systems in the United States and other countries |
|
about faq's links products |
|
Welcome
to Semantic Compaction Systems,
*Registered Trademark of Semantic Compaction Systems in the United States and other countries |
[1] What areMinpseak® brand softwares?
[2] Why do you have to use pictures?
[3] Isn't it hard to learn all of the associations?
[4] Isn't it hard to learn the sequences?
[5] How does a person learn all of this?
[6] How much time does it take to learn all of this?
[7] How do you decide on the codes?
Minspeak® brand
softwares are powerful pictorial systems used in augmentative communication.
They allows fast, accurate access to language and voice for a very
large number of individuals. The Minspeak® approach
to language representation is patented and owned Bruce R. Baker. Over
the past 25 years, Minspeak® language representation
systems have had many contributions by clinicians, special educators,
linguists, and people who rely on augmentative communication. In principle,
Minspeak® systems can work on many different
kinds of computers and handhelds, if the computer is set up to use
this approach to language and language representation. However, though
the Minspeak® system can work on any computer,
a series of computer-based language aids developed by the Prentke
Romich Company (PRC) have been specifically designed
to take advantage of the educational clinical, and communicative power
to represent natural language. The devices which have been specifically
designed for Minspeak® language systems are the
Pathfinder, Vantage, Vanguard/Vanguard II, SpringBoard and ChatBox.
The Minspeak® Application Programs used in these
device are based on the UNITY®128**
language program. **Registered Trademark of Semantic Compaction Systems
in the United States and other countries* In 1980 Bruce Baker had
his first contact with individuals who could not use speech or hand
signs successfully with unfamiliar communication partners. He saw
the communication aids and visual languages (VL's) used at that time
and sensed that they did not really meet the needs of those people
who relied on them for communication. In particular, he noticed that
all the augmented communicators whom he met were significantly slowed
down by the number of "hits" it took to generate speech with spelling.
Just count the number of
letters and spaces in the preceding sentence. The total is 182. If
an individual were using a head stick and generating a letter every
two seconds, it would take 364 seconds or over six minutes. This is
not counting possible mistakes and giving no rest periods. Of course,
not all sentences are this long. Nevertheless, sentences often run
75 or more keystrokes in length. The amount of time and effort spent
in generating language this way is not practical.
Baker became aware of a
variety of approaches to the keystroke problem. Some had suggested
that a linguistically-oriented prediction program could "solve
the problem" of the number of hits. Such a program could operate
in a way that would allow a person to select a letter and then read
a list of words to find whether the word he or she wanted might be
on the list.
This kind of computer-based
system has the potential to reduce the number of keystrokes, but another
problem intervenes. How many of us can remember what we want to say
if after every keystroke we are presented with a list to be read?
How much time would be involved in reading list after list? Our example
sentence would require a system operator to read at least two dozen
lists.
Baker was also aware of,
but rejected abbreviations as a way to lower the number of keystrokes.
The problem here is that an average person uses thousands of words.
The possibility of memorizing hundreds and hundreds of abbreviations
seems both overwhelming and futile. Because there are only 26 letters,
the abbreviations would soon become arbitrary. For example, how many
words begin with "w. " There can only be 26 two letters abbreviations
with "w" and these would include such abbreviations as "wq" and "wx."
Letters, even with computer
power, did not seem to be a viable choice. It is interesting to note
that Baker came to this conclusion even before he learned that the
majority of individuals with developmental disabilities had difficulties
in reading and spelling. He came to this conclusion interacting with
college-educated adults who relied on augmentative communication devices.
He saw that these individuals were not, at that time, nor would they
in the future be, able to achieve voice output communicative competence
using the alphabet.
His own background in ancient
and oriental languages predisposed him to look for semantic or meaning-based
solutions. For instance, a typical Chinese sentence may have a dozen
or fewer Chinese characters. Imagine the benefit, he thought, of a
language representation technique for English that would average under
20 characters per sentence.
Unfortunately, the "logographic"
or Chinese approach creates another significant problem. The Chinese
writing system has thousands of different characters. To achieve communicative
competence using a Chinese-like logographic system, a person would
need at least 1,000 characters. Imagine scanning through such a set!
Baker explored the possibility
of more direct semantic representations, but here a similar problem
arose. A child by the age of five uses thousands of different words.
The possibility of using thousands of different pictures seemed unrealistic
to him Even if a "bare bones" vocabulary could be established, many
hundreds of words would have to be represented. A person who relies
on augmentative communication would be faced with an overlay featuring
hundreds of pictures. This might work with a manual or eye-gaze system
where the teacher or clinician points to areas of the board, but it
is impractical for use independently.
The electronic alternative
to this would be a menuing system where an individual would be forced,
for example, to select a category of words and then have the overlay
change to show just the pictures in that category. While this approach
might be neat for showing the different members of a category to an
augmented communicator, it would not help much with the actual language
as we use it.
For instance, when people
are on the screen for emotion words, they rarely select more than
one emotion to build a sentence. For example, in the sentence, "Many
people like going to the zoo," a system operator would probably find
the "many" picture on a different screen from the "people" picture.
That picture in turn would be on a different screen from the "like"
picture. The "like" picture might be on a different screen from the
"go" picture, The "go" picture might again be on a different screen
from the picture representing "to." And even the "the" picture might
be on a different menu from the "zoo" picture.
If "like" and "go" were
on the same screen and "the" and "zoo" would be on the same screen,
some of the problem involved with multiple screens would go away.
But the solution would create another problem. If operators are going
to select more than one picture from a screen, they need to tell the
machine to leave each screen when they are finished. It can't be done
automatically. So not only do users need to tell the computer which
screen they are going to, but they must tell the system to leave each
screen when they are finished.
Such a system did not seem
reasonable to Baker. He knew that able-bodied people generated speech
effortlessly and semi-automatically. Introducing a complex range of
visual search and discrimination tasks, menu to menu, picture to picture,
and screen to screen, seemed very unnatural to him. Just the time
spent visually refocusing would be a major stumbling block without
considering the level of mental distraction.
Consider a somewhat parallel
situation. Touch typists are able to remain within language without
bothering about reading lists or key placement. If someone introduced
a linguistic word-prediction system or a menu-driven picture system
to a touch typist, keystrokes would be saved, but the results would,
of course, be chaos. Baker was aware of other
difficulties in representing language with a picture for each word,
not the least of these is the problem of synonyms. For instance, how
does a person distinguish among the pictures for "mad," "unhappy,"
"sad," "frown," etc. In English, as in all other languages, these
words are encoded by using sounds which are arbitrarily linked to
the concepts they represent. The sound "sad" encodes the notion "sad"
simply because people who speak English use it that way. In French
"triste" means what English speakers mean by "sad." These words do
not illustrate a concept, they encode it. It is the natural language
capacity of humans to encode rather than to represent meaning directly
that allows people to speak and think automatically.
Because of his linguistic
training, Baker was impelled intuitively first to a pictorial system
because pictures "give more bang to the punch." Such expressions as
"A picture is worth a thousand words" have more than a grain of truth.
The letter at best encodes a phone or phoneme. A picture can encode
an entire idea or even a suite of ideas.
Second, he was drawn to
encoding systems because they allow for rapid, easy (automatic) language
generation. He was also aware of the necessity to have a relatively
small set of symbols. Combining these notions together, he explored
the development of hieroglyphic-like systems because these systems
allow for efficiency, great flexibility, automatic processing, and
a small symbol set. His first efforts closely reflected an interpretation
of hieroglyphic systems in the light of current microchip technology.
The result was the first Minspeak® language representation
system.
Baker's approach was relatively
simple, but it has proven over the years to be practical and quite
powerful. Many people assume that pictures have one meaning. A picture
of a cup means a "cup." A picture of an apple means an "apple." However,
objects do not exist in reality out of context and that context has
been used all along in AAC. On language boards, people use a picture
of a cup to mean things like "drink" and "thirsty." An apple can mean
"fruit," "eat," or even "hungry." What Baker did was to systematize
a natural process. He used the broader meaning of a picture for the
purpose of encoding. For instance, a picture of a rainbow on a manual
board can mean "happy" or even "rain." The interpretation, the meaning
of the picture, is often in the interaction between the augmented
communicator and his or her communication partner. People who have
had experience in interacting with augmented communicators using picture
systems are aware how clever the "point and guess" game can be.
In a Minspeak®
software system, rainbow could mean "happy" or "rain" according to
the context in which it is used. But rather than depending upon the
communication partner to determine the meaning of the picture, Minspeak®
software allows the augmented communicator to select this meaning
independently. A picture of a rainbow sequenced after a picture of
an umbrella can mean "rain," but rainbow sequenced after a key illustrating
a heart (representing emotions) can mean "happy." Rainbow as the first
key can be designated to mean "colors." Rainbow followed by heart
can then mean "red."
Using this simple process,
Baker discovered that he was able to encode hundreds of words and
sentences with very few keystrokes (two or three symbol sequences)
in ways that can lead directly to automatic language generation. Automatic
language generation begins when a symbol sequence becomes so familiar
that the operator no longer thinks of it. At this point, the physical
effort is substantially lessened.
The person now thinks directly
of the word itself, not the symbol sequence. This is the way able-bodied
individuals talk. When they have mastered the motor pattern for producing
a word, they no longer think about that motor pattern. It becomes
automatic. After developing children learn how to produce a word,
they learn through using the word. They do not remain focused on how
to produce the word.
Unlike with other symbol
approaches, the augmented communicator can be allowed to decide on
his or her own meanings. The computer is programmed to reflect the
meanings that the user has decided upon. As with all electronic and
non-electronic communication systems, the process of vocabulary and
picture selection can be a long one. Consequently, early in
the development of Minspeak® system, the idea
arose to make software application programs with picture and vocabulary
selection already done. The first of these application programs was
dubbed Words Strategy® , it encoded more than
2,500 different vocabulary units with an average keystroke of 1.9
hits per word. The Words Strategy® application
program allows an augmented communicator to generate language flexibly
and more than 60% greater than if the same communicator were spelling.
More than 2,000 professional person hours were spent on vocabulary
selection and the selection of the icon sequences for this application
program.
Minspeak® application
programs limits the involvement the communicator has in navigating
from screen to screen, and it avoids having him or her to read list
after list of predicted words. Approximately 100 to 180 hours of learning
on Words Strategy® (currently Unity®
application program), roughly the equivalent time required to master
touch typing, are needed to bring a person to a level of automatic
processing on the system.
Baker's work has generally
focused on adult language, communication rate enhancement, conversational
competence, automatic processing, and other issues dealing with interaction
and fluency. However, speech clinicians, special educators, and other
professionals have taken the system in many different directions.
In particular, Minspeak® application programs
have been found to be a significant tool in teaching language concepts
to people who have developmental disabilities.
Perhaps the best way to
understand a Minspeak® language system is to
look at some examples of its use by several individuals who rely on
augmented communication. In the examples, pictures will often be referred
to as icons. Calling pictures icons on computer systems is a wide
spread practice. The term icon and picture can often be used interchangeably.
Lindsay, who is five years
old, has a red APPLE icon on her Liberator. The APPLE is used to represent
the following words and sentences: "I want to eat," "When are we eating?",
" I brought money for lunch," "eat," "hungry," "red," "food," and
"kitchen." When she presses the APPLE followed by another key, it
will say one of the above words or sentences.
Lindsay's teacher, family,
speech-therapist, and classroom aid all grasp the idea of the APPLE
picture representing many ideas or words, but their question is, "When
Lindsay presses the APPLE picture, how does the Liberator know what
word or sentence she wants to say? "The answer is that the meaning
of the picture is defined by the other pictures used with it.
When Lindsay wants to say
the word "red," she first presses the picture of the RAINBOW, which,
for her, codes the idea of colors and then presses the APPLE. If she
wants to ask "When are we eating?," she first presses the picture
of a QUESTION MARK, which codes the idea of questions, and then the
APPLE. The RAINBOW and question icons narrow down the topic to colors
or questions. For each of Lindsay's messages with the APPLE icon,
she first presses a different icon to code the topic and then presses
the APPLE icon.
In Lindsay's system, she
uses nothing but two icon sequences. The first picture always codes
the topic and the second picture always codes the specific idea she
is communicating. She has 32 pictures on her overlay. Each picture
codes a different topic and she has up to 32 sentences and single
words with each topic. Lindsay's Minspeak® system
is organized very efficiently for her, but that doesn't mean that
all Minspeak® systems are or should be organized
the way Lindsay's is.
If Lindsay's pictures each
had one and only one meaning, she would have just 32 messages in her
system. By using the broader or contextual meaning of each picture
to form codes, Lindsay has a vocabulary of several hundred words,
phrases, and sentences in her system all available from a single overlay.
She does not have to ask someone to change her overlay, and she does
not have to navigate from screen to screen in order to talk. As Lindsay
learns the motor patterns used in each sequence, language generation
becomes more and more automatic. Lindsay can then concentrate more
on using language and less on finding language in her system.
Sara, who is 22 years old
has one, two, and three icon sequences. Using approximately 60 icons,
she has nearly 800 single words and phrases coded according to part
of speech (e.g., noun, verb, adjective, preposition). She presses
one or two icons and then the part of speech needed to code the word
she wants (e.g.,, APPLE + NOUN = food, APPLE + VERB = eat, APPLE +
ADJECTIVE = hungry, APPLE + THERMOMETER + VERB = cook). Although Lindsay's
system is different from Sara's, both people use multimeaning icons
in sequences. Regardless of how the systems are organized, the operating
principle of a Minspeak® brand software remains
the same: The use of multimeaning icons in icon sequences.
The principle that pictures
have multiple meanings and that specific meanings can be defined by
the combination of pictures has been grasped by Lindsay at age 5 and
Sara at age 22. As Lindsay matures and her language needs increase,
her system will gradually become more similar to Sara's. Although
the icons will change and the icon sequences will vary and increase,
the principle in Minspeak® brand systems will
remain the same.
Baker, as a theorist and
linguist, has developed a visual system which has been a powerful
language and communication tool for Lindsay, Sara, and numerous other
augmented speakers. Not only does the use of Minspeak®
language systems reduce the symbol set needed and save keystrokes
when generating language, but it blends naturally with the therapeutic
language needs of experience by many people who rely on augmented
communication.
Lindsay is learning to
create a cognitive map of language. When looking at a picture, she
is increasing her ability to associate a wide range of language concepts
(e.g., What is it?, What do we do with it?, What goes with it?, Where
is it found?, What shape/color/size is it? etc.). This type of language
intervention parallels the same intervention and retrieval techniques
used with speaking children who have language and communication disorders.
One might ask, "Why is
it important to have a visual language system which promotes language
development?" The answer is quite simple. Because the majority of
augmented speakers have significant language and vocabulary deficits.
The visual system used must do more than just represent the bare bones
of language. It must promote growth and development. Anything short
of that is not in the best interest of the person relying on augmented
communication.
In summary, what are Minspeak®
brand systems? They are the use of multimeaning pictures
in picture sequences. They provide a powerful and natural means of
coding language, no matter what natural language one represents..
It is the only visual language system used by augmented communicators
which, simultaneously, is efficient, promotes automatic processing,
and supports, through its own structures, language development.
Some people who use Minspeak®
visual language cannot read; therefore, words are useless
to them. Also, pictures naturally invoke multiple ideas, so more information
can be stored with fewer pictures and less space on the communication
board. For example, the word apple can be used to express the
idea of an apple, whereas the picture of an apple can express not
only the word apple, but also eating, the color red, hungry, New York
(the big apple), worm, etc. Remember, a picture is worth a thousand
words.
The associations of the
pictures to their meanings are designed to be memorable, and make
frequent use of many common mnemonic devices, such as homophony, rhebus,
cultural association, shape, etc. (See Baker et. al in Minspeak®
Conference Proceedings 1990). Also, Minspeak®
visual language systems are designed to be customized, so
a user can assign meaning to pictures according to his or her own
experience, thus making the association more memorable.
The sequences of icons
in a Minspeak® visual language system are just
as memorable as the associations to the pictures, for the same reasons.
(See questions 1 and 2 above)
A person learns all of
this through practice, practice, practice. Just as an able-bodied
child learns to speak using his or her own speech organs through trial
and error and many years of speaking, so does a child with disabilites
learn language and how to use it with his or her own AAC system. People
who can teach Minspeak® language systems can
be professionals such as special educators, speech pathologists and
their facilitators or other persons familiar with the system, such
as parents, friends, fellow classmates, etc. Some people can teach
themselves with the device and manual at their disposal. These people
are usually cognitively intact teens and adults who already know how
to read.
Training time, as with
any learning task, varies from individual to individual. We recommend
approximately 90 hours, for a cognitively and linguistically intact
person, to learn a prestored vocabulary of approximately 2000 words
well enough to carry on a normal everyday conversation. A person can
begin talking immediately after learning a few words and phrases appropriate
for certain situations. They may not be able to compose complete,
grammatically correct sentences, but they will be able to communicate
some of their ideas to others, which is the most important goal of
the speech act.
A Minspeak®
visual language system is completely customiziable. A user
can define all of the codes by himself or herself. However, creating
such a system requires hundreds of hours of intricate detailed work,
since codes rely upon other codes and the semantic networks created
by the pictures and the codes are tightly interwoven. As a result,
Prentke Romich Company (PRC) offers pre-arranged vocabulary sets known
as Minspeak® Application Programs (MAPs). We have found that it is
more effective for an individual to spend the time learning a pre-arranged
vocabulary rather than creating a complete vocabulary for himself
or herself. PRCs
MAPs have been developed by teams of speech/language pathologists,
special educators and linguists with extensive experience working
with individuals with mild to severe physical, cognitive and multiple
impairments.
[Back
to Top of Minspeak® Brand Software FAQ] Simply
Language!
To contact Semantic Compaction Systems, send e-mail
to Bob Conti at minspeak@minspeak.com
©Semantic Compaction Systems
Last update:October 10, 2005