See Abstracting, More on Abstracting, and particularly Consciousness of Abstracting. Also click on search, type in "abstraction" (without the quotes), select general semantics files, and click search.

I understood the circularity of knowledge dashed arrow from verbal levels back to the event level as showing that our language habits affect how we perceive as well as what we project the wigo "to be" - both the Whorfian hypothesis and our constructed model of what we believe exists. We see what we believe until we trip over what we did not believe. (A failed prediction.)

I think that our working is a little different from the abstraction process depicted in the structural differential, which is a one-way up process. But what seems to actually happen is that incoming sensory abstractions in conjunction with what we are doing stimulate the recall of past experience. This is neither abstraction nor projection. It is the neurological equivalent of a metaphorical transformation. What gets recalled is an abstraction from memory, from prior experience, which includes what we did in the past similar situation. The majority of the incoming sensory data gets replaced by our expectation based on past experiences. See updating general semantics. This is one reason we are better at believing than at seeing. The current structural differential makes no allowance for memory recall.

I have been working on an updated representation that takes into consideration the above. It shows the memory recall process and circular feedback within the memory stimulation loop. The dashed arrow is missing, but it is verbally described.

The relation of one level of abstraction to the next is not a "function", as there are many possible ways to abstract from one level to the next. A key notion of function is that it has a singular value on any starting point, but abstraction is a relation that can have many possible values at or on any starting point. In a more general sense, we may say that we don't know what the value of a function is based on a starting point.

Agreed - the character of abstraction changes whenever the input and output mediums differ. From light to sensory responses. From sensory responses to nervous processes. From nervous processes to muscular activity. From muscular activity to sound waves (which we understand as verbal). Each time the medium change, the technology of abstraction changes.

A function is a map from many things in a territory to many map elements such that given one starting point in the territory there is only one resulting element in the map. May different points in the territory can "go to" the same element in the map, but one point in the territory can not go to two or more different points in the map. If the map shows a house, it might cover a whole block of houses in the territory (many points), but the map cannot show two different elements. If it does, then it is not a "function" but only a "relation". We can overly two different functional maps such that, at the same point, each functional map has a different element. An example might be two demographic charts which show average income in one chart and percentage of home ownership in another chart. These are two different functions, each having a single value an any point. There cannot be two different averages in the same location.

Functions have only one value at any given point.
Relations may have multiple values at any given point.

Correct, but the mathematical notion of function is very specific, it can have only one value. The outcome depends on the starting point and the nature of the function. Abstraction by just one level from a starting point of N pieces of information can have as many as 2^N possible outcomes, including none and all. Taking into consideration the fact that these N pieces of information can stimulate many possible memories, the result can be many more possibilities. For many people, the result does prove to be "functional" in that they can only come up with "the" answer. Consciousness of abstraction allows one to recall that there are many possible results, and delayed reactions allows us to think of many more. That could be the feedback with different instances of abstraction, but overall, the result is a relation rather than a function.

I "quibble" on the use of the term 'function'. Even in general usage it should be and usually is thought of as producing only one value. A function is dependable in that it gives the same result for the same inputs. If something does not do so, it is not a function, but only a relation. For example correlations and probabilities are not function relations, because the results are many to many. A function is many to one.

Korzybski used the notion and the notion of propositional function in the same way.

An abstraction relation also changes one abstraction into another, but it allows for more than one possible result, whereas a function implies a single result.

Since you mentioned mathematics, and I'm trained in mathematics, we should be on the same page.

Regards,
Ralph

Ben,
Thomas's use of single quotes is academic standard. The Institute's Bulletin editors chose to do exactly the opposite of academic standard. See On the use of Quotation Marks, Etc.: A Review of General Semantics, Vol. 51, No 1, Spring 1994.

The general semantics community should drop their parochial and cultish attitude about quotations and adopt academic standards. Double quotes for quotations and "scare" quotes, single quotes to mention a term such as the word 'word'.

It is well known in philosophy and language that quoted words are what is known as an opaque context, and that means that the normal meaning of the term may not apply. That is essentially what "scare" quotes mean. I urge the institute to make a decision to adopt this standard usage. Single quotes are only for when one is talking about the quoted words directly (or for quotes within a quote).

Gee, Ben,
You asked about the use of single quotation marks in the first place, and you followed up the answer with your reply that you used them differently (in the non-standard way), so the sub-topic of the use of quotation marks had three prior entries in this topic before I chimed in. If you would like to move the four postings prior to your administrative interjection, starting with your first inquiry, to a new topic say "quotation marks", it is your prerogative as moderator to do so. If you choose to do so, your latest off topic post and this reply would no longer be relevant, so they both could be deleted.

It seems I've got a number of questions to respond to.

Steve wrote, Ralph, one of the things I like about GS is that I find the material generally pretty easy to understand and yet has been important to my personal development. Some of the material in the links above makes for tough reading and is tougher to assimilate. Any chance we can get you to refine that down to something more accessible?

I don't think that can be said for the majority of Science and Sanity.

Abstracting is a technical study using context specific definitions intended to provide some precise structural notions.
More on Abstracting shares a conversation contrasting perspectives relevant to various aspects of abstracting. It is not easy reading, as the contrasting perspectives differ in point of view as well as language style.
Consciousness of Abstracting is specifically intended to illustrate the fact that between any two verbal levels of abstraction is an entire sequence of abstraction by the person processing the first level of abstraction. People seem to think that verbal abstraction between levels is as simple as lexical, logical, and semantic abstraction. Any computer, with well formed rules, can perform these types of abstraction. Between any two verbal levels of abstraction, the abstractor involved engages in the steps depicted in neural abstraction, so, the process "ain't so simple"; To "dumb it down" involves too much abstraction, so much so, that the gist of the article will be lost. It behooves anyone who want to consider himself a general semanticist to look at the full complexity as I have outlined it. Brains and language are never as simple as they have being made out to be. To make an analogy with a road map, I can't get from my home to your using ONLY a large area map of the United States that shows only major highways. I need the detailed local area street maps in the beginning and in the end. Using the structural differential as a model of the human abstracting process is as futile. -- I just discovered that Bois, in The Art of Awareness (second edition) on page79 shows "abstracting through [semantic reaction] filter" between each of Korzybski's levels, including between any two verbal levels.

Steve wrote: This does enlighten my sense of the importance of the "downward" aspect of abstracting. I still wonder if the "dotted line" from top to bottom in the SD accounts for this, at least in terms of Semantic reactions, but new to me to consider the abstracting process itself as color by our word-category projections (1) and Ralph, I think the criticism to GS for the lack of the "downward" aspect to the abstracting process may be a fault/lack in the Structural Differential, but as you indicate K includes the Whorf-Safir hypothesis in his discussions. Here is a discussion of that Sapir-Whorf-Korzybski (2).
The first thing to note is that neurological abstracting does not have a direct "downward" direction. What happens, according to On Intelligence, is that the brain keep a running picture of what we are doing and where we are. It is constantly looking at its map of prior experiences and anticipating or predicting what it expects to see next. A huge amount of incoming information, which includes what we are doing (our motor responses) at the moment, predicts exactly what to see next. As we open the door to the bathroom at home, we expect to see the toilet exactly where we saw it before when we were stepping in and turning our head just so, and flicking the light switch on. None of this information is a "surprise", so it is relegated to low-level "ignore when what we see matches our prediction". If the toilet were missing from its expected spot, that prediction failure will be flagged and "sent upstairs" to the conscious processing levels with a "red flag" saying "big problem"; we low-level peon processing areas need your upper management help.

This process, which I might label as anticipatory feedback, governs what incoming sensory abstractions reach higher levels in the processing. All memories are non-elementalistically stored as a whole combination of sensory experiences, motor activity at the time, a record of our hormonal state, our level of hunger, how tired we are, etc., along with what we have been doing before this point. If some reward is being experienced, then the nervous system actually takes the time to retrace, backwards (back propagation), the nerve circuits of recent activities leading up to the reward. This strengthens the synaptic connections, resulting in giving some higher priority to remembering (and re-using) the sequence that led up to the reward. So we have anticipatory feedback and back-propagations that both "sound like" they might be some kind of "downward abstraction", but that would be to mis-describe the process - create a bad map. Abstraction is one-way - up. It can be by minuscule steps, even so far as simply repeating the lower level, but that might be called "improper" as in the whole set being an "improper" subset". To borrow from set theory terms, "proper" abstraction would not mean a simple copy from one level to another. In as much as the structural differential shows only an upward direction for abstracting, it is correct.

We've talked about the Sapir-Whorf hypothesis, in which it is hypothesized that our words affect what we perceive. The brain research and mechanisms uncovered show that the anticipatory process, that acts at many levels, is "projecting" downward what the various levels of neurological processing are expecting to see next. I describe this as abstracting into "pre-existing categories". When the expectations are not met, they are "kicked upstairs" to the next level for more processing. The next level may handle the failure, such as seeing a towel draped across the toilet. We know what to do. Move it, because we have many experiences that generalize to moving an obstacle to gain access to what the obstacle blocked. We learned this non-verbally as a child. Not much consciousness required. Not very abstract. Piaget called this process assimilation. When the discrepancy between what is expected and what is seen propagates all the way up the neurological hierarchy until it reaches our full consciousness (and for general semanticists, consciousness of abstracting), it may require altering our map. Piaget called this process accommodation. Here's a good explanation. The Whorfian hypothesis holds that we actually perceive or see things according to the assimilation process. Korzybski included the dashed arrow from the verbal label level pointing back to the event level to depict this.

When, however, people refuse to accommodate new data, they are using their map, their expectations, their predictions, etc., in spite of evidence to the contrary. This behavior is what we call intensional orientation. It can be as simple as not seeing what you are looking for because you forgot what color it is and are looking for something the wrong color. It can be as stubborn as a southern redneck or a religious zealot. When we seek, we are enhancing or focusing on the predictive factor, sometimes so strongly as to wash out the sensory input. The result is we only see what we believe is there. ("I already looked here, and it's not here.") This "abstraction failure" occurs pretty commonly... ("But I already looked there, and it wasn't there!").

Knowledge of recent brain function allows us to relate general semantics principles and processes to what is going on in the brain as we currently understand it.

The Whorfian hypothesis is implemented in the assimilation and neurological prediction mechanisms. It is shown in the structural differential as the dashed arrow.

Abstraction works only upward, but which higher level possibilities are chosen depends strongly on the "downward" prediction (what's expected) and the history of back-propagation of reward for successful action (our learned experience).

General semantics treats the kind of accommodation failure that gets labeled "intensional orientation" as a "bad thing". Recall however, that Occam's Razor states do not multiply essentials beyond necessity, so we need a balance between accommodation and assimilation. We don't want to move in the direction of too little structure by making changes too often. We also don't want to move in the direction of too much structure (hardening of the categories) by failing to update our maps when necessary. A duck cannot "unlearn" its imprint on the first moving object as its mother. But that particular "intensional orientation" has survival value for ducks, as the first moving object it sees usually is its mother.

Abstraction is always "up" into pre-existing categories until such time as an accumulation of evidence that does not fit requires revising the category structure. Prior to that time we always "see" in terms of the current category structure (the downward projection ala Whorf), both non-verbal and verbal. This also addresses Steve's question.

Steve, My copy of The Art of Awareness is the second edition, and it does not have the diagram you asked about, so Bois must have dropped it when he updated.

Milton wrote, I will again suggest that all general semantics principles can be experienced (extensional). If we don't make an effort to experience these experiences, we can get bogged down in definitions; we might give more importance, more significance to our definitions than to what these definitions are about. (intensional. We won't be practicing general semantics. We won't be using the tools so much as talking about and defining them.

Milton, I do not see these as either-or (two valued) or as contrasting dimensional directions (infinity valued, but opposed). I think one can be using the general semantic principles while carefully experiencing and applying them to verbal levels such as formulating and reformulating the principles, including developing updated definitions.

An "extensional orientation" in such research would include searching and checking the literature, and revising one's draft statements accordingly before offering them publicly. Consciousness of abstracting becomes very important in reviewing formulations from the point of view of estimating what the writer did with his sources, as well as considering alternative ways of interpreting what we ourselves are reading, knowing that our interpretations depend on our own experiences, etc..

We can talk about the tools while using them; we can be using the tools while we talk about them. We can be experiencing the tools while analyzing and revising formulations of them.

While your general caution is reasonable, it might be more valuable if you would point to specific instances (extensional) where you think it is a problem. In a discussion forum such as this, the formulations "are" the primary topic; examples of how to experience each formulation is part and parcel of our topic area. Do you think we need more examples here?

I don't assume that my views and understanding of the past have not changed.

Since characteristics "cross" from one level of abstraction to another, the term itself is being used in a multiordinal sense at both levels of abstraction. If I think in terms of stimulus-response, I avoid the single word 'characteristic'. The Whorfian factor takes the word characteristic and the string drawing showing a connection, and maps it to the notion of some thing "going" from one level to another, something, again a singular term, connoting an invariance of some kind. A common "string" of invariance. A transduction and a transformation takes place such that the response bears no similarity to the stimulus. But this flies in the face of those who claim our abstractions must be similar in structure to the territory.

How could we visually represent abstraction of characteristics (level x) to characteristics (level x+1) so that they do not look like invariances?

Something like this?
.
.
v
v
+
+

My idea is to not use the same word (characteristic) at two adjacent levels of abstraction, and to draw something that changes between the two levels. Even indexing the word characteristic doesn't seem to be enough difference. That's why I suggested periods changing to plus signs. Any linear icon with a break symbol (represented by the two v's) to connect to a different linear icon.
If it were horizontal I could use something like:
event---------->>============object
object---------->>============verbal

I don't get the dog example.

I can experience getting bitten, and that's an object level response, and I can scream, "That darned dog bit me", and that's a verbal level response.

A sophisticated understanding of the structural differential by bringing to it many different experiences with various aspects of general semantics together with developed consciousness of abstracting allows one to remember that the simple picture of the structural differential is overly abstract, and that we must supply the indexes and differentiation "in our consciousness" as we look at it and in our speech as we talk about it. These "meanings" and semantic reactions are in the nervous system of the beholder. The beginner and novice general semantics initiate does not have these developed capabilities, and our visual memories provide the cues to recalling the "teachings". If it isn't in the map, we have a tendency not to be reminded, even knowing that the map is not the territory. Look at all the discussion of characteristics and multiordinal. The simple suggestion of changing the strings from a homogeneous representation to a two-phase heterogeneous representation would show these in the simple map. I also suggest that we change the representation of characteristics to: miniature broken parabolas in the event level, miniature circles in the object level, and miniature label shapes in the verbal levels. Doing this will greatly enhance the beginner and novice ability to assimilate "characteristics" as multiordinal and different from level to level. The label to label strings should also be two-phase.

Recall the adage:
If I hear, I forget;
if I see, I remember;
but if I do, then I understand.

So a beginner trainee explaining the structural differential as modified will understand when he or she talks about the miniature shapes - different at each level - as "characteristics", and the shape differences will remind him of the multiordinal character of "characteristics".

But, like has happened at the many seminars I attended in the past, explanations of the existing structural differential, even by the "experts", has neglected to mention that "characteristics" are different at each level of abstraction. According to my recollections, they have even been explained at such seminars as representing "invariances" from level to level, and this is part of the teaching that that explained how our mapping produces maps with "structure similar to the territory" - by "relative invariance under [abstraction] transformation" from level to level going "down physically" (up in abstraction level) the structural differential levels. It is a nice neat simplification, but it introduces overly simplistic errors that result is subsequent difficulty with understanding the multiordinal nature of "characteristics".

Milton, the term "consciousness of abstracting" in the context you cite was the title of a link to an article. It was not being used to mean the process of consciousness of abstracting. Did you click through and read the article? Consciousness of abstracting dictates that we try to be aware of all levels of abstracting, including the neurological depicted in the article.

Is it my imagination, or has this topic switched from "abstraction" to "multiordinality"? I think, in ordering the complexity of general semantics notions, "multiordinality" depends on "abstraction", but not the other way around.

Abstraction is certainly a multilevel or multistage process going from one level to the next. I have provided examples of several kinds in my article Consciousness of Abstracting prior to applying those types to "consciousness of abstracting".

Milton, The article ends with, "But let not all these possibilities paralyze us. Go forth and act, but be prepared."

What's the "opposite" "complexify"? "Dumb down"?
We must remember that abstracting between two verbal levels involves semantic reaction, and the structural sequence of a semantic reaction is as depicted in the first list in the neural abstracting section. Moreover, I did not expand "7. Collections of neurons that selectively respond to different combinations of sensory cell responses." to explain that this also means recalling memories and experiences.

I submit that the majority of people who participate on this list are not simply asking questions for so-called "experts" to answer. This list involves discussions of the finer points and subtler interpretations, including how to update general semantics.

Perhaps the learn-gs topics should be divided into beginner, intermediate, and advanced discussions, such as introduction to general semantics, novice interpretations, and expert discussion group - or some other multi-level classification to put the simple stuff in one place and the esoteric stuff in another place.

It is also the job of the moderators to determine what is and is not appropriate for this list and for any particular topic area. We can suggest to them that "learning gs topics" be broken up into two or three sub-group areas such as frequently asked questions and answers for beginners, advanced general semantics topics discussions, and discussions on updating general semantics. That way you can have the "simplifications" in the beginner area while I "play" in the other two areas. Any answers provided in the beginner area must be consistently stated Doctrine and simple sound bites.

What do you all think? Do you think it would be beneficial to organize "learn gs topics" into two or three different levels of abstraction depending on complexity - or some other criteria to differentiate the target audience by level of experience?

Ben, The continuation has been posted here.

The continuation is now here.