LC3cA61 Knowledge of Objects: Backwards Events Analysis
I recall discussions with Mimi Sinclair about the extensions of the object concept:
she proposed I focus on the exploring the infant’s understanding of “inclusion within concave objects.” She had been doing related work with Stanback in Paris, and thus could directly propose a set of objects affording the opportunity for experiences which could permit the infant to learn about inclusion in concave objects — and that this would serve for about three years, because usually children resolved these issues around their third birthday. She suggested I make a collection of objects to be used only in the experiments, and proposed collections of nesting cups, and some balls that fit in some of them, and nesting boxes, and some blocks that fit in some of them, and some rods, of differing lengths. We made this the basic set of “standard objects” used in our “experiments.” These were supplemented at various times by other objects, a “Ring Tower” puzzle, everyday objects used functionally in her life, such as a hairbrush and her cup and spoon. Of course, over time Peggy accumulated many toys, which were always welcomed with her.
What about the features of these objects? Some were largely irrelevant, such as color, and others were more “distractors” than features, such as pictures on the bottoms of the nesting cups (e.g. giraffe, baby carriage, etc.). It is very hard to specify what Peggy saw, or what she made of what she saw. What we CAN specify is her use of these objects in her actions. As for any infant, oral exploration was the universal: the infant put everything in her mouth, whether moving to put her mouth on the object or, as later, bringing the object to her mouth. What was she doing? And is it not an interesting puzzle that children do not understand inclusion, when their primary early sense organ, the mouth, operates by inclusion?
Here is how I approach this problem
We begin with the assumption that what the child is focused on is what she uses in her own actions. The questions she is trying to resolve with her actions might be “does this produce milk?” or “does it taste good?” I see no hope for resolving such questions with the material available to me. But if she is asking “will this fit in my mouth?” one might explore the issue of “mouth-eye coordination” from observations in this corpus. If an object is “mouth-able” that is an early size discrimination which does not depend on visual coordination, but could produce an accumulation of experiences which eventually will aid mouth-eye coordination.
We can look at Mimi Sinclair’s advice, which she said was not a “theory to be disproven” but “une sondage,” an exploration to take a look. There is a circularity here: the features which matter are those that the subject uses; what she uses are those features that lead to results of her actions which interest her. How can we tell about these issues? By looking at the video clips tracking threads of behavior which lead to discriminations and control, a kind of regular success in achieving some end. Does not such a proposal raise immediately the question of intention? Absolutely! And here is how we approach it. It is obvious that some objects afford opportunities for action — a ball will roll in two dimensions, but a cylindrical rod will roll only in one. Does an infant make that distinction initially and then conduct an experiment to see whether Object X will roll in two or only in one direction? No, she brings it to her mouth, finds whether it is small enough to fit and may even notice that a rod fits one way but not another, then she throws it to see what happens. Peggy did it often. How do we connect such explorations of what develops to ways we think of things?
In a succinct summary of the labors of science, Charles Peirce said there are three kinds of valuable work. First is determining the values of constants. Second is the formulation of general laws. Finally, most importantly for Pierce, was discovery of new phenomena. What I see in this corpus is that the infant discovers phenomena (new to her) as she creates for herself the hallucinations in which everyone of us lives. So, in discussing these matters with adults, we need descriptions which make sense to the adults we all have become, even though everyone has abandoned the genetic-epistemological thread which led to this mature view. You and I think in terms of “properties” of objects; that is wonderful, effective and embodies multitudes of discriminations. We want to know when an infant appreciates a property and how that came to happen. But there is more: we want to know about relationships of properties between objects. For example, if you place a solid object on a larger concave object, the solid object could disappear or, if square, could stay on top unless rotated to align corners. Let us say that regular success requires a grasp of the relationship. Further, one might ask if there a general rule — as the flawless solution of the Ring Tower puzzle requires object application in order by decreasing inner diameter sizes. With Nesting Cups, the integration of different collections in separate partial orders requires re-evaluation of order in the entire ensemble (or might it not? Analysis will tell.) In this case, regular success requires understanding of a general rule. How can we even talk about these issues?
Go back to Mimi Sinclair’s proposal of what objects to make part of our standard set.
We can see the mature mind appreciating the challenges the infant faces. Does that help us? Yes, in setting out a field of exploration. But there is a second question in respect of Intention. Which uses are their own ends and which can best be understood as actions serving another goal. For example: when Peggy throws an object, is it for the satisfaction of throwing something (occasionally surely so) or is it an experiment (e.g. to see if it will roll). Of course, an action can be both at once. Can one distinguish such cases (e.g. throwing different sorts of objects proves this is employed as a technique to create an interesting result… I think one will be able to argue one can, on the basis of the individual situations examined; but that is an issue to be decided by the data). Time to bring into high contrast what adult views expect of object interaction and what Peggy actually did. I can do it and intend to.
The Interpretation Framework: how to do it
Euclid especially knew one needed to understand Quod Erat Demonstrandum (what must be proved) — and create predecessor structures and arguments from which one could reach THAT conclusion by lucid deduction (this is the mathematics of “Backwards Analysis“). What will we do?
– find, late in the study, clear evidence of regular (repeated, expected) success.
– work backward through the corpus to find earlier examples where there was no such regular success; this could be of different types, such as:
> failure: when a clear goal was blocked, with evidence of non-repairability
> incompletion: a clear goal was blocked, then dropped.
> opportunities to pursue a goal not undertaken at all: for this situation to obtain, we would have to see that Peggy “appreciates” a goal without “accepting the challenge” the goal implies. To say Peggy can “appreciate” a goal is too abstract; one needs to use a different word for this than appreciation. Of the synonyms I find, most useful might be to “detect” a goal, because it is non-committal in respect of just how specific any “appreciation” might be. Such a “detected” goal would mean that Peggy both perceives and values some end state that she has not brought about previously.
> accidental discovery: this could be one key type of goal discovery process; think of using the rings to keep balls from rolling way, and the discovery that, when the ring is lifted, some balls go with it and some do not; this is the discovery of a new phenomenon, albeit a very concrete one.
> observed success of others: think here of Peggy seeing me bring out a stack of her nesting cups all in a compact form; then dumping them out; is this not a challenge to her to get them back together?
> instruction: someone specifically demonstrates a phenomenon to her; the question is whether she “makes anything of it” or not. how do I resolve that issue? Consider a “grown-up” analogous problem. Having seen a Rubik’s cube in its “solved state,” then seeing a disordered one, you might wonder how to fix the cube. You could “detect” an interesting goal without addressing it at all. You might see others fix the cube or hear them talk about it. With an accumulation of such incidents there may come a point when you decide that you should be able fix a cube yourself, and thereby adopt that as a goal yourself. Eventually, mastery could be within your grasp. (If you want to master Rubik’s cube, see my discussion here, published in July 2013.)
In sum, We will proceed with our analyses to identify points of significant mastery, then proceeding backwards through the corpus seek plausible examples of adoption of the goal; subsequently we will attempt to identify the behavior segments in the period between adoption and mastery and then mine those segments for information about how the infant’s learning was making progress. Of course, it will also be important to identify episodes relevant to goal detection before the goal had been adopted, and to examine them for evidence about possible experiences leading to goal adoption. The intention is to examine exhaustively all pertinent material in the intervals before adoption and before mastery to discover predecessor actions which become components ultimately integrated into the regular successes of mastery.
For a concrete example of how the analysis will proceed generally, refer to LC3cA62 Analysis: Mastery of the Ring Tower Puzzle.
The grand analytical game is to get from contexts to concepts — from interactions with the surround — via recognition of opportunities and exploratory objectives to intentions, then to formulated goals, and to debuggable plans.
For more detail on the grounds of analysis, refer to LC3cA61a Describing Objects and Actions
and to LC3cA61b Inferring Goals from Affordances and Outcomes
Return to LC3c-Analysis.