In the first part we left a child of two and a half with a red circle in front of them and two cards to choose from: a red square and a blue circle. The instructor said “pick the same one” and the child picked the blue circle. We saw that there is no bare “the same” — somebody decides the same in what — and that Julio Varela answers “have they learned it?” by changing one piece of the situation and watching whether the correct answers hold.
Emilio Ribes asks an earlier question: when that child responds to what is on the table, what kind of contact are they having with it?
4. Ribes’ question: what kind of contact is this?
Ribes does not ask how much survives a change. He asks what kind of relation the child is having with what is in front of them. He tells five apart, and each has its own way of being done well.
The fastest way to see them is a kitchen.
Figure 6. A kitchen in five numbered panels, same child and same instructor: (1) the lights come on and the child looks up; (2) the child turns the tap and water runs; (3) two glasses, one fuller, and the child points at the fuller one; (4) the instructor, with her back to the cupboard, says “the blue cup is on the top shelf” and the child fetches it without her pointing; (5) the two sitting and talking, with a speech bubble that says “why did you say ‘up’?”.
- Coupling. The lights come on and the child looks. Their response adjusts to something that happens, but changes nothing. What is asked of them is to tell apart (Ribes calls this differentiality).
- Alteration. The child turns the tap and water runs. Their action produces a change in the environment; the world responds to what they do (effectiveness).
- Comparison. Two glasses; “which one is fuller?”. The answer is not in either glass on its own: it is in the relation between the two, and that relation changes if I change the glasses (precision).
- Extension. The instructor describes where the cup is, and the child acts on something they cannot see, guided by what they were told. Language is needed, and another person is needed (congruence: what is said and what is done coincide).
- Transformation. Talking about how something was said: “why ‘up’?”. It is language about language (coherence).
Each of those five scenes is a single illustrative moment, not a demonstration. Looking up once when the lights come on does not establish differentiality, and asking “why ‘up’?” once does not establish transformation on its own — Ribes’ own criteria for each contact call for repeated, varied instances, which is exactly what an index needs and a kitchen story cannot show.
Two things worth knowing about this list, so as not to turn it into another ladder:
- It is not a scale of difficulty, nor a single measure. Each contact has its own criterion of adjustment. A child can be perfectly adjusted in a coupling; that is not “worse” than a comparison, it is a different thing. In Interlaza, the Ribes module scores three of the five — coupling, alteration, comparison — with the index Ribes published for each, never with a shared share of correct answers. Extension and transformation have no published index for their criteria (congruence, coherence), so the module leaves those two unscored rather than inventing a number.
- The last two need language, and the fourth needs another person. Ribes tried to set up extension alone, on a screen, with students aged 16 to 18 (2009): none of them acquired it. That is why Interlaza does not promise those two contacts with two-year-olds — a caution that applies to every age-based inference in this article, including this one, since the studies behind these tests were run with older children and adolescents, not toddlers.
5. Ribes’ warning: being right does not say which contact you are in
And now the sentence of Ribes’ that has changed most how we read a result. A matching task, he says, “from a formal point of view favors the possibility of comparative contacts”; but “from a functional point of view” it can be solved “through coupling contacts”.
Translated: the table of the first part has the form of a comparison. But if the child always sees the same pair (red circle → red circle), they can get every trial right by repeating, without comparing anything. Recognizing the same object, Ribes says, “does not constitute comparative behavior”: it is coupling. And a fixed but arbitrary pairing (if red, then yellow) is still coupling, because it never changes either.
Figure 7. Two children, each with a results sheet reading “10 out of 10”. Above the first, a thought: “this card goes with this card”. Above the second: “the one that is the same IN WHAT IS IN CHARGE today”. Below, the line: “Same result. Different contact. The sheet does not tell them apart.”
When is there comparison in Ribes’ sense? When the properties vary and the relation holds: today color is in charge, tomorrow size, the sample changes, the positions change, and the child is still right because they respond to the rule in force and not to the cards. Ribes adds one more condition: there must be at least two checks against each other; a single pair compared is not comparison.
That is what separates the two children of Figure 7, and it is what the app has to be able to separate too.
6. Where the two meet
Up to here they look like two different questions. Ribes joined them himself.
In 2009, asking how to know which contact a child who gets a matching task right is in, he wrote that the available indicator is Varela’s transfer tests: if the correct answers survive a change of modality, there is comparison; if they survive a change of relation, there is extension; if they survive a change of dimension, there is transformation. And he added that those indicators “are not definitive”.
Figure 8. A bridge with three arches. On the left bank, “Varela: what survives the change?”. On the right, “Ribes: which contact is it?”. Over each arch, a correspondence: “change the modality → comparison”, “change the relation → extension”, “change the dimension → transformation”. Under the bridge, in small type: “Ribes, Vargas, Luna & Martínez, 2009”.
So the answer to Varela’s question is the method for answering Ribes’. One gives the criterion for what is going on; the other gives the test to find out. Neither settles for the share of correct answers.
The same study left two results that affect us too:
- Training the “previous” contact was neither necessary nor sufficient to acquire the next one. The idea that the simple must be mastered first and the complex afterwards, as a compulsory ladder, did not hold. In Interlaza the prerequisite hierarchy (identity, then different exemplars, then the word) warns but never blocks.
- A group that only watched the trials, with nobody telling them whether they were right, learned the first three contacts. Seeing the arrangement teaches something on its own. It is one of the reasons the first phase of Interlaza’s errorless teaching puts the correct answer beside the sample and barely lets the child choose: first the relation is seen, then it is chosen.
7. What Interlaza does with the two questions
Four concrete things. Each is a decision of the app that can be checked.
a) It checks what it taught, without correcting while it checks. At the end of a curriculum block, where the block’s own content supports it, there is a probe: a handful of trials in which one piece changes (a new exemplar, or the property in charge), without telling the child whether they were right. It is Varela’s method. Not every block gets one — a block whose transformation has no name that fits it is left without a probe rather than given one that would measure the wrong thing. The result is stored apart from the training and is never mixed with it in one percentage.
b) It alternates the criterion and shows it at the top. In the alternating-criterion activity, one trial asks for “same color” and the next for “same size”, with the same cards. At the top appears a pair that exemplifies the rule (two things of the same color, or two things of the same size), built for each trial and never repeating the sample or the answer. It is Ribes’ permutation with Varela’s second order, and it is how Interlaza tells the two children of Figure 7 apart.
c) It names what changes when you leave the tablet. A probe with a real object after learning with photos is not “a change of medium” and nothing more. In Varela’s vocabulary three things change at once: the instance (photo → object), the modality (weight, texture and volume appear) and the mode of the response (looking and touching → grasping). And because stimulus and response change together, it is transfer in the most demanding sense, not simple generalization.
Figure 9. The tablet with the photo of a spoon on the left; a real spoon in the child’s hand on the right; between them, three labels on the arrow: “another exemplar”, “other properties (weight, texture)”, “another response (grasping)”. Below: “Three changes at once. That is why it is measured, not assumed.”
d) It classifies the task, never the child. The app can say “this activity, as it is set up, is coupling” or “this probe changes the modality”. It does not say “this child is at such-and-such contact”. That could only be known with the child’s whole history, and Ribes himself reminds us that the same exercise is a different contact for two children with different histories.
8. What we do not claim
So that the article does not promise more than the app:
- There is no scale of difficulty. Varela’s fifteen types are a classification; Ribes’ five contacts are five different questions. Neither list is walked in order.
- A “transfer” probe does not demonstrate transfer by being called one. It demonstrates what came out: what was held constant, what changed, what the child did.
- Position and order appear here as something that is controlled, not something that is taught. The correct answer moves around so that “always the one on the left” does not score. Teaching a child to compare by order — what comes first, what is closer to the start — is a different task, and this article does not cover it.
- The Varela and Ribes studies we cite are with children aged 9 to 11 and with adolescents. They serve to define the tests well, not as direct evidence of effectiveness with two-year-olds.
- A failed probe is a question about the material, not a verdict on the child. The app’s response is always about the teaching: more variety of exemplars, or check again later.
Glossary in three steps
Each term, in the order the article itself asks for: example → explanation → name.
- The light comes on and they look → responding to something without changing it → coupling.
- They turn the tap and water runs → producing a change → alteration.
- “Which one is fuller?” → responding to the relation between two → comparison.
- They fetch the cup described to them without seeing it → acting on what was said → extension.
- “Why did you say ‘up’?” → talking about talking → transformation.
- Being right with cards they never saw → coming loose from the concrete → functional detachment.
- Being right after something changed, with nobody saying whether they were → transfer probe.
Sources: Ribes-Iñesta, E. (2018). El estudio científico de la conducta individual. Manual Moderno, ch. 10 and p. 370. — Ribes, E., Vargas, I., Luna, D. & Martínez, C. (2009). Adquisición y transferencia de una discriminación condicional en una secuencia de cinco criterios distintos de ajuste funcional. Acta Comportamentalia, 17(3), 299-331. — Varela, J., Padilla, M. A., Cabrera, F., Mayoral, A., Fuentes, T. & Linares, G. (2001). Cinco tipos de transferencia. Revista Mexicana de Análisis de la Conducta, 27(3), 363-383. — Kantor, J. R. (1958). Interbehavioral Psychology. Principia Press. Full references are on our science page.