Caring for Other People

Over time, a couple of things can happen:

How can we best support them?

Start Fuzzy?

Initial limitations on the resolution of the information streams is argued (not proven) to be beneficial to subsequent development stages; for example, in computational models:

You do have to evaluate computational modeling with some caution when it is used to support speculations about what is happening with people, but I do prefer it ethically to methods that require the training of captive animals (e.g., the research on prenatal distinction of phonemes discussed below), or even worse, invasive testing on animals.

On the outskirts are invasive but relatively harmless methods used on animals, such as the research conducted on the auditory environment in sheep, the results of which (by the way) suggest that the prenatal auditory stream might not be as muted as claimed elsewhere.

Computational models will always be suspect because they don’t measure humans (and code ultimately only does what you program it to do… after it is done, of course, pissing you off by not doing what you simply want it to do as you debug^⇲); similarly, the early research on prenatal auditory processing needs to be taken with a judicious pile of salt, as they are only extrapolations from postnatal observations of such trained animals as chinchillas and quail.

But, again, appealing to more direct alternatives would cross ethical boundaries.

So what follows is based on an assumption that (a) the set of research that I am allowing to be considered is (b) good enough to entertain this discussion. If you say, “But invasive surgical procedures on unwitting victims proves beyond a doubt that you are utterly wrong about everything,” then I’ll do no more than respond, “Ok then, I’m simply wrong, given your chosen foundations. No problem. Now there is no need to continue this little talk.”

For those of you, however, who are not inclined (in such a manner) to wring personal benefit from contemptible sources of victimization, we can continue…^⇲

Start Small

It is not terribly controversial to contend the following: when a blasé, run-of-the-mill, “normal” person is born (i.e., of the type briefly described above), they have already been developing the skills that will be required to organize their perceptions. They already enjoy at least some of the skills required to distinguish repeated patterns in streams of connected speech (which do not tend to contain much intrinsic noise).

Early in your prenatal career (or perhaps avocation), you start out being able to distinguish the sounds used in any language spoken in the world,^⇲ but then the language environment in which you swim will impose its peculiar pattern of exposure, and that pattern provides the basis for an organization that (naturally) reflects an eventual preference for the system of sounds that have been reaching you.^⇲

We then deliberately help babies figure things out by strictly bounding (i.e., segmenting) our speech for them. We repeat canonical words in isolation from more complicated streams of speech, saying simple things like, ‘mama’, ‘papa’, ‘poop’, ‘dammit’, and so on.

But it doesn’t seem to take very long before we go right back to overwhelming the hapless waifs:

We stop bothering to be careful in part because so very many infants seem to do well enough with this task even when we don’t help them out very much. ^⇲

But we are not concerned here with the majority of people. Some among us could continue to benefit from that sort of help for a longer period of time because their development is delayed.^⇲

The Overwhelm

Of course, it’s not just the auditory time-stream that newly-minted humans have to deal with. Speech is nice in that it tends to be a simple stream with natural pauses for breath. Visual input, for example, is not nearly so kind.

It takes entire books just to describe how a human learns to distinguish figures from their backgrounds, much less come to the realization that a certain set of visual images all add up to the same object being rotated and/or translated in spacetime.

Remember earlier, when we talked about putting boundaries on forms (so that they can be linked with conceptual structures)? Well, all of the following properties contribute to forms:

We are phenomenally lucky that sound treats most of us as well as it does.

Let’s stick with vision for the moment as a representative member of the rest of these bastards.

Illustrative Illustrations

When it comes to clarifying what is meant by time-series segmentation, it just so happens that literal illustrations provide nice illustrative examples (albeit strongly constrained specifically to visual processing), such as the awareness that:

Creating a gestalt is a successful use of a skill, it is not a strategy that compensates for a failure to see the day in the park, or of failing to construct the puzzle picture.

If someone were unable to resolve the entire picture due to an inability to have handled segments at a fine enough degree of resolution (i.e., to have an awareness of the parts needed to form the gestalt), then they might try to resolve boundaries of lower resolution features (which we will get into below), such as masses of light and dark (or the entire phrase between pauses for breath).

So part of Starry Night might be “the thing with the swirly bits” (except felt as a whole feeling, not as those separate words). Note that the picture frame would be helpful for segmentation in that it would present a boundary, increasing the triggers for change-point detection… which, although a technical term, is what you probably would guess it is, namely figuring out whether/when the behavior at a point in a time-series has changed enough (compared to the behavior around it) to qualify as a relative difference (give or take).

Some among you^⇲ might ask whether you’d have to resolve the whole picture to form a gestalt, or if it would count if you only resolved a subset of the information into a smaller whole. Well…

Component Gestalts

Yes, seeing one figure in the painting, or finishing one figure in the puzzle,^⇲ would each be an example of successful gestalt formation (if/when^⇲ things work that way). It does involve figure-ground differentiation, and other types of perceptual organization. It does lead to the creation of a unified whole.

Note that these “local” gestalts arise after the person has successfully segmented their component parts at a fine enough resolution, such as the dots in the painting, or patches of light and dark.

Those local gestalts are what might get built along the way, serving as components of a subsequent, higher-order (“global”) gestalt.

Higher-order gestalts based on one type of segmentation (such as the dots) would not be evoked by components that had been based on a different type of unit (such as the light/dark patches).

While this is still not what the ~GLPs are doing early on, the evidence seems to suggest that it would not be inconsistent with an intermediate step in developing their skills.

Change Points

When someone who is challenged by segmentation is able to resolve any chunk of an information time-stream, something has (generally) acted as a helper to define boundaries in the environment at that level of resolution.

Part of it might be standing at enough of a distance that the individual dots are not distinguishable.

Or maybe there is a figure that is delimited in particularly stark contrast against the ground (increasing the intensity of the change point).

Maybe it’s a corner of the puzzle.

Or maybe – just maybe – it’s a phrase (an auditory and/or visuomanual stream) that happens to be naturally bounded at either end by silence (which is often determined by comfortable breathing patterns); in other words, this would be just like what folks have been observing with the ~GLP hallmark utterances.

So, your intuition about there being something whole, even if it’s not a gestalt?

This is it.

It’s an environmental chunk that is naturally segmented by significant change points.

YAE!^⇲

Here is another imperfect metaphor that might be helpful (nonetheless) in illustrating the notion of change points.

When someone (a) treats their food as an undifferentiated mass (i.e., an unbounded environmental time-stream), and (b) ingests it in portions that are (helpfully) segmented only by (c) the change point that is defined by the size of their mouth (but unfortunately not by the rest of their alimentary tract), we do not then (d) identify them as “gestalt food processors.” We do not treat their lack of portioning as if it were a successful instance of a food unifying function, or praise them for engaging in an alternative type of eating.

It’s not that we want to withhold praise.^⇲

It’s that we don’t want to camouflage a risk.^⇲ That type of untruth would undermine a person’s ability to make informed decisions about their health care.

We recognize that this difference in eating style risks harm, and we support strategies that will help to keep a person safe as they develop more conventional methods of eating.

We help them learn to exercise graded control over the segmentation of their food-stream into portions that will not overwhelm their system.

We often start by cutting up their food for them, establishing distinct boundaries between the portions that constitute safe bites; in such cases, we are the something that helps to establish the change points that bound their segments.

Now imagine that what we are talking about is the way in which someone deals with the massive amount of information that their environment feeds to them.

And now understand that language is a largely serial stream of information (albeit with concurrent suprasegmental encoding in series parallel), where some people have particular difficulty in knowing (a) where the container ends and the food begins, and (b) what the size of a bite is supposed to be.

Q: When do speech streams have natural pauses?

A: In the places where breath forces it.

Q: What if the natural boundaries are not enough to keep a person safe?

A: Then we add some for them while they are learning.

Artificial Augmentation of Environmental Segmentation

In special education, we know enough to reduce the environmental noise for people whose information signal is polluted by an undifferentiated sensory mass.

We have break rooms that are lower in stimulation, and areas in which we divide signal generators into distinct locations (i.e., swings, lights, fans, and so on).^⇲ We make sure that those generators can be individually (de)activated.

Sometimes we know enough to do the same for the general information mass that is assaulting the person. Their sensory organs help to resolve some of that, structurally sorting the reception of light from sound and so on, but even that is not always a help when sensation is transduced and perceived and so on (as in synaesthesia, ideasthesia, or whatever it is that “sensory processing disorders” actually turn out to be^⇲).

Compensation

Sometimes, we know enough to say:

instead of…

But for nearly all of their life (or lives, perhaps), the language stream coming at a challenged segmenter will not have been specially prepared; for example, and generally speaking, the TV really couldn't care any less than it does.

When a naturally-bounded segment occurs prominently in the environment of someone who is having difficulty with this skill, it is no small wonder that they would glom onto it as it whips by (usually again); similarly, it is likely that they will associate that form with some sort of coincident conceptual structure.

Such a person is lucky if they can not only (a) entrench a breath-delimited form such as, “Did everyone remember their towel,” but also then (b) link that segment with a conventionally germane conceptual structure like time to go swimming, as opposed to their desire to be presented with food.

Of course they’re going to repeat that segment when they want to go swimming (or when they are hungry). That survival strategy has to compensate until they use it often enough to have a hope of breaking it into its parts.

It’s a life preserver.

Consistency

So let’s also try sticking with consistent vocabulary.

Don’t treat all of the following as equally interchangeable synonyms, namely: mad, angry, upset, anxious, frustrated, escalated, agitated, doyouneedtotakeabreak. At least in the early stages of the challenge, pick one… maybe the one that the SLP has made prominent in their AAC system (generally in consultation with the family and the autism and/or behavior specialists, and so on).

It’s like when a pet can’t figure out what their name is, because everyone keeps calling it by a zillion terms: bunny, rabbit, fuzzbutt, lettuce, jumper, cutey pie, pellet eater, thumper, my little doll, owdontbiteme, and so on.

It is more important for your choice to be (a) consistent (both in its sound/sign form and its link to a predictable meaning), and (b) bounded on either end by silence, than it is for it to be (c) super short.

And I’m Not So Sure about You^⇲

That doesn’t even take into account the complication that natural languages throw into the mix for all of their users; for example, I would like you to determine the number of words in each of these examples:

cellphone

coworker

breathtaking

everyday

livingroom

shrinkwrap

roundtrip

longterm

ribcage

wellbeing

lifesize

fulltime

forever

moreso

cornfed

login

Compounds exist because the boundaries are soft, so for any given language the answers will change with community (and across time).

Now how about these conventionalized units:

People vary in how they resolve these (and significantly longer utterances).

And of course, all of these answers vary among languages (e.g., the infamous Kraftfahrzeug-Haftpflichtversicherung), where the equivalents of many of the above words are not analyzed into parts.

Now imagine if everything were like that for you.

Entire sentences.

Whole paragraphs.

Where would you start?

What kind of support would you like?

Would it help or hurt you to be identified as someone who processes those unresolved language masses as gestalts, when you are not doing that?^⇲

Caring for Other People

Start Fuzzy?

Initial limitations on the resolution of the information streams is argued (not proven) to be beneficial to subsequent development stages; for example, in computational models:

On the outskirts are invasive but relatively harmless methods used on animals, such as the research conducted on the auditory environment in sheep, the results of which (by the way) suggest that the prenatal auditory stream might not be as muted as claimed elsewhere.

But, again, appealing to more direct alternatives would cross ethical boundaries.

For those of you, however, who are not inclined (in such a manner) to wring personal benefit from contemptible sources of victimization, we can continue…⇲

Start Small

We then deliberately help babies figure things out by strictly bounding (i.e., segmenting) our speech for them. We repeat canonical words in isolation from more complicated streams of speech, saying simple things like, ‘mama’, ‘papa’, ‘poop’, ‘dammit’, and so on.

But it doesn’t seem to take very long before we go right back to overwhelming the hapless waifs:

We stop bothering to be careful in part because so very many infants seem to do well enough with this task even when we don’t help them out very much. ⇲

But we are not concerned here with the majority of people. Some among us could continue to benefit from that sort of help for a longer period of time because their development is delayed.⇲

The Overwhelm

Of course, it’s not just the auditory time-stream that newly-minted humans have to deal with. Speech is nice in that it tends to be a simple stream with natural pauses for breath. Visual input, for example, is not nearly so kind.

It takes entire books just to describe how a human learns to distinguish figures from their backgrounds, much less come to the realization that a certain set of visual images all add up to the same object being rotated and/or translated in spacetime.

Remember earlier, when we talked about putting boundaries on forms (so that they can be linked with conceptual structures)? Well, all of the following properties contribute to forms:

We are phenomenally lucky that sound treats most of us as well as it does.

Let’s stick with vision for the moment as a representative member of the rest of these bastards.

Illustrative Illustrations

When it comes to clarifying what is meant by time-series segmentation, it just so happens that literal illustrations provide nice illustrative examples (albeit strongly constrained specifically to visual processing), such as the awareness that:

Creating a gestalt is a successful use of a skill, it is not a strategy that compensates for a failure to see the day in the park, or of failing to construct the puzzle picture.

Some among you⇲ might ask whether you’d have to resolve the whole picture to form a gestalt, or if it would count if you only resolved a subset of the information into a smaller whole. Well…

Component Gestalts

Note that these “local” gestalts arise after the person has successfully segmented their component parts at a fine enough resolution, such as the dots in the painting, or patches of light and dark.

Those local gestalts are what might get built along the way, serving as components of a subsequent, higher-order (“global”) gestalt.

Higher-order gestalts based on one type of segmentation (such as the dots) would not be evoked by components that had been based on a different type of unit (such as the light/dark patches).

While this is still not what the ~GLPs are doing early on, the evidence seems to suggest that it would not be inconsistent with an intermediate step in developing their skills.

Change Points

When someone who is challenged by segmentation is able to resolve any chunk of an information time-stream, something has (generally) acted as a helper to define boundaries in the environment at that level of resolution.

Part of it might be standing at enough of a distance that the individual dots are not distinguishable.

Or maybe there is a figure that is delimited in particularly stark contrast against the ground (increasing the intensity of the change point).

Maybe it’s a corner of the puzzle.

So, your intuition about there being something whole, even if it’s not a gestalt?

This is it.

It’s an environmental chunk that is naturally segmented by significant change points.

YAE!⇲

Here is another imperfect metaphor that might be helpful (nonetheless) in illustrating the notion of change points.

It’s not that we want to withhold praise.⇲

It’s that we don’t want to camouflage a risk.⇲ That type of untruth would undermine a person’s ability to make informed decisions about their health care.

We recognize that this difference in eating style risks harm, and we support strategies that will help to keep a person safe as they develop more conventional methods of eating.

We help them learn to exercise graded control over the segmentation of their food-stream into portions that will not overwhelm their system.

We often start by cutting up their food for them, establishing distinct boundaries between the portions that constitute safe bites; in such cases, we are the something that helps to establish the change points that bound their segments.

Now imagine that what we are talking about is the way in which someone deals with the massive amount of information that their environment feeds to them.

Q: When do speech streams have natural pauses?

A: In the places where breath forces it.

Q: What if the natural boundaries are not enough to keep a person safe?

A: Then we add some for them while they are learning.

Artificial Augmentation of Environmental Segmentation

In special education, we know enough to reduce the environmental noise for people whose information signal is polluted by an undifferentiated sensory mass.

We have break rooms that are lower in stimulation, and areas in which we divide signal generators into distinct locations (i.e., swings, lights, fans, and so on).⇲ We make sure that those generators can be individually (de)activated.

Compensation

Sometimes, we know enough to say:

instead of…

But for nearly all of their life (or lives, perhaps), the language stream coming at a challenged segmenter will not have been specially prepared; for example, and generally speaking, the TV really couldn't care any less than it does.

Such a person is lucky if they can not only (a) entrench a breath-delimited form such as, “Did everyone remember their towel,” but also then (b) link that segment with a conventionally germane conceptual structure like time to go swimming, as opposed to their desire to be presented with food.

Of course they’re going to repeat that segment when they want to go swimming (or when they are hungry). That survival strategy has to compensate until they use it often enough to have a hope of breaking it into its parts.

It’s a life preserver.

Consistency

So let’s also try sticking with consistent vocabulary.

It’s like when a pet can’t figure out what their name is, because everyone keeps calling it by a zillion terms: bunny, rabbit, fuzzbutt, lettuce, jumper, cutey pie, pellet eater, thumper, my little doll, owdontbiteme, and so on.

It is more important for your choice to be (a) consistent (both in its sound/sign form and its link to a predictable meaning), and (b) bounded on either end by silence, than it is for it to be (c) super short.

And I’m Not So Sure about You⇲

That doesn’t even take into account the complication that natural languages throw into the mix for all of their users; for example, I would like you to determine the number of words in each of these examples:

cellphone

coworker

breathtaking

everyday

livingroom

shrinkwrap

roundtrip

longterm

ribcage

wellbeing

lifesize

fulltime

forever

moreso

cornfed

login

Compounds exist because the boundaries are soft, so for any given language the answers will change with community (and across time).

Now how about these conventionalized units:

For those of you, however, who are not inclined (in such a manner) to wring personal benefit from contemptible sources of victimization, we can continue…^⇲

We stop bothering to be careful in part because so very many infants seem to do well enough with this task even when we don’t help them out very much. ^⇲

But we are not concerned here with the majority of people. Some among us could continue to benefit from that sort of help for a longer period of time because their development is delayed.^⇲

Some among you^⇲ might ask whether you’d have to resolve the whole picture to form a gestalt, or if it would count if you only resolved a subset of the information into a smaller whole. Well…

YAE!^⇲

It’s not that we want to withhold praise.^⇲

It’s that we don’t want to camouflage a risk.^⇲ That type of untruth would undermine a person’s ability to make informed decisions about their health care.

We have break rooms that are lower in stimulation, and areas in which we divide signal generators into distinct locations (i.e., swings, lights, fans, and so on).^⇲ We make sure that those generators can be individually (de)activated.

And I’m Not So Sure about You^⇲

Would it help or hurt you to be identified as someone who processes those unresolved language masses as gestalts, when you are not doing that?^⇲