The Application of Carl Hempel's Epistemological Program to Daniel Stern's Research

by Jerome Liss, M.D.

P.zza S.M. Liberatrice, 18, 00153 Rome, Italy
Tel.: 39 - 06 - 5744.903

Part I: The Need for an Epistemological Protocol: The Problem of Neglecting Concrete Reality

The purpose of this article is to develop a valid and rigorous protocol in order to help guide the processes of scientific observation and theory construction in psychology. Theoretical development in psychology, especially in clinical psychology, has often claimed to be "scientific," but critical analysis shows such a claim to be unsupported and even erroneous.(Farrell, 1986). Starting out with "psychoanalytic investigation," for example, the Freudian theoretical matrix has been adhered to by numerous schools in clinical psychology, sometimes with criteria of "orthodoxy", that suggest a certain analogy with religious orthodoxy. And even today, one hundred years after such an inauspicious beginning, books continue to be written in clinical psychology overstuffed with theoretical elaborations that can leave the reader confused and even stupefied. "What is this author saying?" The theoretical jargon is so abstruse that the reader can either fall into a cloudy-minded trance state, or else close the book feeling frustrated or indifferent. Why is that? Because in these presumably "scientific works", nothing is said that connects to concrete reality, and thus, the basic premise of scientific thinking -- thoughts connection to perceived reality -- is entirely disregarded. What is our method? In this article I will first present certain epistemological principles offered by the known philosopher of science, Carl Hempel, (1952) and then apply these principles to a scientific work of special interest, that of Daniel Stern (1985).

Carl Hempel
Who is Carl Hempel? Along with Otto Neurath, Rudolf Carnap, Charles Morris and other philosophers of science, Carl Hempel developed the International Encyclopedia of Unified Science, a series of works that has served as the cornerstone of scientific theory construction since the 1940's. (Editor, Otto Neurath) To my mind, Carl Hempel's contribution offers a set of fundamental principles that every research psychologist can profit from. I will draw from his classical work, Fundamentals of Concept Formation in Empirical Science. (Hempel, 1952)

What do we mean by epistemology. "Epistemology" means the knowledge we have about our knowledge. Another way to put it: "How do we know what we think we know?" What are the criteria according to which we can say that our knowledge is more or less valid? I am avoiding the phrase "true knowldge" because, as we shall see, a scientific epistemology cannot tell what us is "true", but can only lead us to assert that certain ideas are more or less "significant", "coherent" and "supported." No more than that. This may discourage the person who seeks to know "the truth," but it also invites those who love constant challenge to enter the turbulent sea of scientific inquiry. Thus our "scientific adventure".

Science Means Ideas Linked to Experience The fundamental principle of science is that it is a relationship between ideas and concrete experience. Prof. Hempel states, "Science is ultimately intended to systemize the data of our experience, and this is possible only if scientific principles -- that is, our theoretical constructions -- have a bearing upon, and thus are conceptually connected with, statements reporting what has been established by immediate experience."(p.21) To put it in a nutshell: When we base our ideas on concrete experiences, we have science. Without concrete experience, no science. Thus, to repeat our earlier critique: Scientific books and articles in psychology, especially clinical psychology, frequently present little or no evidence, that is, little or no reference to concrete experiences. Thus, they are not scientific. To be scientific, we must link ideas to experience. To cite our original source, Carl Hempel: "A theoretical system without empirical interpretations is incapable of test and thus cannot consistute a theory of empirical phenomena. We shall say of its terms as well as of its concepts that they lack empirical import."(p.39) However, this fundamental epistemological principle of what science is all about represents, as we shall now see, not only a requirement that is rigorous, but also an opportunity.


Design I: What Is Science? IDEAS (interaction) OBSERVATIONS: Overcoming the Traditional Paradigm

In an earlier article, "The Philosophy of Science and the Clinical Researcher," (Liss, 1988) I suggested that this fundamental principle of scientific epistemology -- to link ideas to experience -- can guide the clinical researcher as he uses his case studies as the basis for scientific theory construction. This previous article indicates that the difficult and nearly impossible task that certain scientific researchers in clinical psychology think is required, namely, that all scientists, even those in clinical psychology, must adhere to the natural science paradigm of creating experimental and control groups, is actually a false requirement for clinical science. Before explaining why this is so, let us first review very briefly what is involved in the natural science paradigm that requires the creation of experimental and control groups. This paradigm stipulates that the competent scientist must create experimental groups which contain the variable, "X", and in which consequence "Y" is measured, and then must make a comparison of these "Y" results to the results of a matched control group which does not contain variable "X". Very neat and logical, but, we propose, only necessary for the physical sciences! The above cited article attemps to free the clinical researcher from the fetters of this unwieldy and inappropriate scientific paradigm, however useful that paradigm has been and remains to be for the physical sciences of chemistry and physics. This more fundamental concept of what science is all about, that is, the epistemological principle that science means the to-and-fro movement between observation and theory construction, permits the clinical researcher to profit from his experience with "single, unique and unrepeated" events -- such as the clinical encounter -- and to use this experience of a "single, unique and unrepeated event" as the observational basis of clinical science.1  But, as this article points out, the job is not an easy one. For where do we put our efforts in order to obtain scientific rigor?2 In the first place, by means of the adequacy of our registered data. (Post-session note-taking is hardly adequate. Too much is forgotten.) Full note-taking duringthe clinical session, (thus, done by an observer) or else, more appropriately, tape-recording (adequate for verbal registration but lacking the visual field) and now, most important of all, video camera recording, represent the primary methods today for obtaining adequate registered observations. These methods, and especially the last mentioned video camera, permit the motivated clinical researcher to satisfy the bottom-line of a scientific project, an adequate registration of raw data.

An Epistemological Map
Following this, the clinical researcher begins his concept formation and theory construction. This is where Prof. Hempel's clarifications pave the way: A scientific theory might be likened to a complex spatial network: Its terms are represented by the knots, while the threads connecting the latter correspond, in part, to the definitions and, in part, to the fundamental and derivative hypotheses included in the theory. The whole system floats, as it were, above the plane of observation. By virtue of these interpretive connections, the network can function as a scientific theory: From certain observational data, we may ascend, via an interpretive string, to some point in the theoretical network, thence proceed, via definitions and hypotheses, to other points, from which another interpretive string permits a descent to the plane of observation. (p.36) Let's clarify this wonderful constrution with a design.

Design II: Carl Hempel's Map of Science

The basic point is that certain theoretical concepts must be rooted in observations. Without such a connection, the theoretical schema can float off the ground without any ties to reality. Such reality-tied concepts are called "operational concepts," an excellent term since this suggests the idea that we are required to "operate," that is, to "act upon" our field of study. And "act" upon it how?  By observing our field of study directly and creating from that observation some form of stable registration. In a later section we will offer some reflections upon this special problem of "registration" as it relates to the clinical experience. The second point is that theoretical concepts are linked to one another through "definition" and "hypotheses." It is this inter-linking among concepts, the theoretical elaboration, that gives us our understanding in depth and complexity and justifies our efforts in scientific theory building. What are we talking about? The above is the abstract schema offered by Prof. Hempel that can help us guide our scientific inquiry. I propose that the best way, however, "to know what we are talking about," is to adhere to the scientific method itself and examine a concrete example. What example?

Daniel Stern
I will use the work of Daniel Stern, previously Professor of Psychiatry and head of the Laboratory of Developmental Processes at the Cornell University Medical Center, and now at the University of Geneva. In fact, I will focus on Chapters 6 and 7 of Prof. Stern's book, The Interpersonal World of the Child, (1985) since the research and theoretical elaboration that he presents in this particular section of the book can serve as an excellent basis for studying the two-edged process of science: observation and theory building. 2A We will now enter into the heart of our argument, which we will also call "the scientific adventure."  


Part II: Intersubjectivity: "Intersubjectivity": A High Order Concept

Prof. Stern begins with a theoretical concept that is not an operational concept but, rather, a concept of a higher order, "higher" in the sense of more general and abstract: "intersubjectivity." He first offers a definition of this higher order concept, "intersubjectivity", and, in order to accomplish this task, he draws upon a definition already employed by another research team, C. Trevarthan and P. Hubley, (1978). "Intersubjectivity" is defined as "a deliberately sought sharing of experiences about events and things." (p.128) The above definition illustrates two further epistemological principles: 1. The theoretical matrix, composed of definitions and hypotheses, as stated by Hempel in his image of a network composed of strings and knots, (above), is already undergoing systematic elaboration. 2. The scientist must not be isolated. To use definitions and hypotheses from colleagues' work is a way to integrate one's own theoretical constructions with those of others. Thus, to profit from another's useful definition is not to shun responsability, as a scientist, but the very opposite: to integrate one's work with the work of others in the same scientific community and thereby widen the interlacing and "scope" of one's project. (The term, "scope," will be made more precise at a later point.)

Design III: From General Concepts to Operational Concepts General Concept INTERSUBJECTIVITY Analytic Reduction Operational Concepts SHARING SHARING SHARING ATTENTION INTENTION AFFECT

An Analytic Reduction
By means of another type of definition, what might be called an "analytic definition," in that it goes from a whole into its constituent parts, Daniel Stern breaks down his general notion of "intersubjectivity" into three more precise types of "sought sharing of experience": shared joint attention, shared intentions, and shared affective states. (p.128) We call this an "analytic definition" because it involves no empirical work, but it prepares us for empiricial work, because it breaks down the whole into a series of constituent parts which are more concrete and therefore closer to the observable plane. The scientist is often engaged in this epistemological effort in order that his general notions become operationalised. These three types of sharing -- of attention, intention and affect states -- will then become Stern's bottom line operational concepts.

Rooting Our Concepts in Observed Reality
Prof. Stern then raises the scientific question of the connection to reality: "What behaviors do infants show to suggest that they can conduct or appreciate these sharings?" One can only admire the elegance of Prof. Stern's step-by-step scientific protocol: from general notion, "intersubjectivity," to specific operational concepts coming from an analytic definition -- the three types of sharing experiences the infant seeks to obtain -- to the bottom-bottom line of the seen behaviors that root the concepts in observed reality. And here are the seen behaviors: When one person "points" in a direction with the hand or a finger and the other person follows the visual line indicated by that pointing, the "interpretation" (a Hempelian word) is that this represents a desire, on the part of both people, to share a common field of attention. "Joint attention." But the observation is not finished. Stern notes that after an infant (even at nine months of age) follows the direction of the mother's pointing finger and then stops and momentarily holds the eye movement (and thus, interpeted, "reaches the target"), the infant then looks back at the mother, interpreted as "validating" that the joint attention has been achieved. In fact the infant often shows a series of back-and-forth movements between the target and the mother's face. This is very important! An operational concept, of course, may be based on one specific observation (i.e., in this case, following the direction of the pointed finger and stopping on the target), but when we have two or more observations that are linked in sequence (the infant's eye movement travelling between the target and the mother's face), and thus correspond to the theoretical interpretation (the infant thereby validating the experience of "shared attention") we are building up our scientific ground by digging deeper into reality.

Interlacing Our Ideas With Those of Colleagues
Sharing Intentions: For this example of "intersubjectivity," Stern cites the work of Bates (1979). What is of special interest here is that Daniel Stern's concept, "sharing intentions," is worded in a slightly different way from J. Bates' concept, "intentional communication." We have already heard Stern (p. 3) Now here is what Bates says: "Intentional communication is signaling behavior in which the sender is aware, a priori, of the effect that the signal will have on his listener, and he persists in that behavior until the effect is obtained or failure is clearly indicated. The behavioral evidence that permits us to infer the presence of communicative intentions includes (1) alternation in eye gaze contact between the goal and the intended listeners, (2) augmentations, additions, and substitution of signals until the goal has been obtained, and (3) changes in the form of the signal towards abbreviated and/or exaggerated patterns that are appropriate only for achieving a communicative goal.

Design IV: The Operational Concept, the Observed Behavior and the Experiential Interpretation

We have, in this instance, an interesting challenge in scientific theory building. Let's look into the matter more closely, because perhaps we can draw out elements that will be frequently profitable when theorists interlace their own theories, with their particular language, with colleagues' theories whose language is similar, but not identical. How can we put together the two theorists' notions, that is, Stern's "sharing intentions" and Bates' "intentional communication?" We will undertake a comparison of the two notions by means of a more precise analysis of their meanings ("a reduction analysis") in order to indicate where there is overlap and where there is a difference of significance. (The following procedure corresponds to Carl Hempel's proposal of "reduction analysis."(pp.25-29) Hempel also cites the epistemological work of Rudolph Carnap (1936) for a more thorough formulation of the term, "reduction analysis", and explanation of the method.) As a note to the reader: The following section of "reduction analysis" and the step-by-step comparison of the analyzed components may become tedious reading. Nevertheless, it is this type of precise and step-wise procedure by which separate scientific theories, proposed by different researchers, can find their exact interlacing.  And, to repeat, we are fortunate to have the availability of Daniel Stern's work, because his research and writing show a degree of clarity and rigor that actually permit a "reduction analysis". In subsequent writings, we will treat other "scientific works" in psychology and, as we do so, we will run into ambiguities, contradictions and confusions that will only make us wish, in this scientific navigation, to return to Sternian waters.  

An Analytic Breakdown to Integrate Concepts  
I propose that if we analyze Daniel Stern's concept, "sharing intentions," we will discover that it includes, implicitly, three components:

In other words, an analytical break-down of "sharing intentions" includes two experiences, having an intention and wishing to share it, and one behavior, that of sharing. We will return to this analytic break-down in a moment to derive other implications.  

Design V: Making More Precise Our Operational Concepts

How is this analytic breakdown of Stern's "sharing intentions" connected to Bates' concept of "intentional communication". Our reduction analysis brings us to the conclusion that Bates' concept is included within that of Stern's; that is to say, "intentional communication" (Bates) is one aspect of "shared intentions" (Stern).  As Bates says, "Intentional communication is signalling behavior."  Therefore, it is an element (3)(above) within our Sternian-based analytic schema, that is, the category dealing with each party's expression of his or her "intention to share". "Intentional communication", which Bates also names "signalling behavior," means the expression of the experienced intention by one party. And so it is a part of Stern's "sharing intentions". But there is still further overlap between the two theorists' proposals: Bates adds, in his very first sentence, "in which the sender is aware, a priori, of the effect that the signal will have on his listener." So here we have, in association with the expressive component of Bates' concept, the experiential component, "Aware of the effect that the signal will have" is a part of the experienced intention. The last part of that same very first sentence, "he persists in that behavior until the effect is obtained or failure is clearly indicated," means a combination of behavior (persisted in) and experience (until the effect is obtained). Thus, a reduction analysis of terms shows us that all of Bates' concepts fall within Stern's definition of "sharing intentions." We might say that a part of the joy and fruitfulness of science comes from "correct sharing" among colleagues. Therefore Prof. Stern is correct to cite Bates' concept of "intentional communication" as a means of elaborating his own concept of "sharing intentions."  In addition, Stern then correctly cites Bates' three criteria of observation (Nos. 1 to 3 in the above Bates citation) as observational criteria also valid for his own concept of "sharing intentions." The point is that science moves ahead when fellow researchers talk to each other and share. The epistemological point is that "we have to know what we are talking about" in order that such sharing (in this case, regarding similarities, but in other cases, regarding differences) becomes, operationally, a helpful process that buttresses scientific progress.

Design VI: Comparison of Stern's "Sharing Intentions" and Bates' "Intentional Communication"


Part III: Four Outcomes of Reduction Analysis More Specific Operational Concepts

Let us now examine four additional outcomes of this reduction analysis and its consequence of permetting us to compare the concepts of "sharing intentions" and "intentional communication." The first outcome is that a seemingly bottom line operational concept, in this case, "sharing intentions," can be analyzed, at times, into yet more precise conceptual categories. That means that our evaluation, at any moment within the construction of a scientific theory, regarding which operational concepts are most basic, that is, which concepts are most closely tied to our observationed data, must remain flexible. We must always admit that a reduction analysis of our ideas can permit more specific concepts to emerge, that is, concepts which show an even stricter linkage to the observational plane.

Avoid Unwieldy Over-Definition
Secondly, I would like to hypothesize that while reduction analysis to the highest degree possible is usually useful for research and theory elaboration, such a precise analysis may not be so useful for the clinical worker. To jump to another kind of example: Let's say that you visit my home and ask me where the bathroom is!  Shall I simply point to the corridor, mention which door, and add that the light switch is in the corridor to the left of the door? Or shall I show you a blue-print of the whole house, its plumbing (for the water) andelectrical wiring (for the light) in order to indicate where the bathroom is? This example is meant to show that over-definition can be excessive and perhaps even act as an obstacle for good clinical thinking in which it is not so important to have a blue-print of the situation, but instead, to have workable concepts that help the clinician get the job done. But this hypothesis will need substantiation.

Distinguish Invisible Experience from Observable Behavior
The third consequence is fundamental! It is best to distinguish behavioral from experiential aspects of a concept. Only the behavioral aspects are directly observable. (The expressive component of "sharing intentions" as well as of "intentional communication", means the behavioral component.) The experiential component is not seen. In this sense, clinical research in psychology always contains a mysterious domain: the experience! As R.D. Laing points out so lucidly in The Politics of Experience, (1967) "I can see your behavior, but your experience is invisible to me." And the converse is equally true: "You can see my behavior, (and that also means hear my spoken words), but my experience remains invisible to you." 3 The various consequences of this fundamental gap between observed behavior and invisible experience, at the interpersonal level, has vast importance, not only for clinical research, but for psychotherapy and for the understanding of all psychological interactions among human beings.4

Operational Concepts as a Synthesis of the Seen and Unseen
And the fourth consequence brings us to a further clarification of what we mean by the basic term in epistemological theory, "the operational concept." Stern's concept of "intentional sharing," (and so, of all operational concepts in science) shows a certain linkage to the observational plane. But "concept" and "observation" are not identical. This is what Hempel means when he says that the theoretical network is "anchored" to the observational plane. In fact, we can always augment and modify our anchorings to observed reality as we develop our operational concepts with further scientific work (a mixture of theoretical work and observational work). The point is evident: The operational concept is not equal to its anchorings in reality but only shows more (and sometimes, through critical analysis, less) numerous and evident connections. Therefore, I can say I know my perception is true, while I cannot say that my ideas --the ideas that I also know-- are true!  Rather, the ideas I know are more or less adequate and supported. Perhaps less evident but of yet greater significance: What is an operational concept? Yes, it is an idea linked to observables. But its significance also comes from the fact that it contains meanings which are not apparently observable (in this case, the experiential component of "sharing intentions.") Thus, it would be appropriate that the theorist, in his theory construction, delineate which aspects (or meanings) of an operational concept are linked to the observed data and which aspects are, at least for the moment, invisible. Do all aspects of an operational concept need to be, in the final analysis, visible, at least as a potential? The "narrower view of empiricism," which Prof. Hempel, as almost all epistemologists of today, finds unacceptable, does say "Yes" to this excessive and narrowing demand. On the other hand, today's more "open position" of empirical epistemology, which Prof. Hempel espouses, says that our operational concepts, as well as higher order concepts, may contain aspects that are invisible, unobservable, not only for today, but for always. The word "experience" may fall into this category. For example, the experiential component of the infant's desire to "share intentions" -- the experience that lies, in a sense, "behind" the behavior, or is it "underneath"? -- is not seen. One may argue that although this is the case for the infant, the mother, on the other hand, can give introspective reports regarding her own experience. So her experience can take on some form of visibility. But the counter-argument is that while some aspects of the mother's experience can become visible or manifested through her introspective reports, such verbal reports, and even drawings, non-verbal behaviors, sounds, and so on, cannot reveal all aspects of "experience": What about our sense of time, sensation of space, the interrelationship among different perceptual modalities and their relationship to both verbal and non-verbal thoughts? The "psychic shape" of our thoughts and feelings? Or "the space" in experience that comes on like a mood of emptiness, fullness, and so on? And what happens within the empty spaces of experience? Our words and our other manifestations can hardly transmit the presence of these experiential components. But there will always remain some "invisible," non-denotational components that are "experienced" but not manifested, especially when they constitute a part of the "peripheral ground" of our experience, to use a Gestalt psychology concept, rather than a part of the "central figure." Therefore, the concept of "experience" a concept essential for most psychological understanding -- will always retain aspects that are invisible, and this, we propose, justifies recourse to the "open view" of empiricism and the need to renounce the earlier "narrow view." The upshot is this: Our operational concepts are concepts that combine immediately observable aspects (the expressed behaviors) and non-observable aspects, (some for the moment, some for always) such as in our notion of experience. But these non-observable aspects are essential parts of the concept's meaning. Some readers might consider that the above considerations are a deviation from the main argument regarding epistemological principles that are required for theory construction in psychology. However, I think that the opposite is the case: The scientific psychologist, whether researcher or clinician, will always have to wrestle with this "ontological fact", namely, that experience has, and always will have, components that remain invisible, notwithstanding our attempts to make manifest, with scientific ardour, more and more of that mysterious domain.

A Concept can be supported by Various Pieces of Evidence
To conclude with Daniel Stern's treatment of "sharing intentions," he once again refers to a colleague's work.(p.131) This time Stern takes his observational material from the researh of J. Dore (1975): A mother is holding in her outstretched hand something that the infant presumably wants, a cookie. The infant reaches out his own hand, with palm up, toward the mother, and looks back and forth between the mother's hand with the cookie in it and the mother's face. At the same time the infant makes grasping movements with his own hand and sounds out, "Eh! Eh!" with an imperative tone. Thus, mother and child are "sharing intentionality"; the mother is intending to let the child obtain the cookie and the child is intending to receive it. Do these observations by Dore correspond to Bates' outline (see above) of the evidential criteria that justify "shared intentionality" and, therefore, is Daniel Stern correct in citing Dore's description?  Yes! to all of these questions. J. Dore's description corresponds to Bates' observational point No. 1, "alternations in eye gaze contact between the goal and the intended listener." Dore's description corresponds, as well, to Bates' observational point No. 2., namely, the elicitation of "additional" signals: the infant's grasping movements and the sounding of "Eh! Eh!". We do not have evidence for Bates' observational point No. 3, a change in the signal pattern, but this type of observation, if I understand Bates' notion, depends upon seeing how the signal pattern evolves over time when there are repeated instances of a similar situation. In addition -- and this point deepens our awareness of how to use operational concepts, an operational concept need not (and usually cannot) satisfy all possibilities of evidential support. In other words, I propose that if we were to enumerate all of the possible maniafestations of our operational concept, we might develop a fairly extensive list. Scientific work means to define several key evidential manifestations of a concept that we are investigating and show the presence of several, not all, within our observational field. Other researchers can collaborate in our scientific adventure, adding further data from different possibilities within the observational plane, perhaps to support our thesis, perhaps to contradict it. More positive evidence, more support for the concept; less positive evidence, less support.  And contradictory evidence? This feeds the conceptual war. Stern then calls up other evidential phenomena to support the operational concept of "sharing intentions": For example, teasing between siblings, private jokes between sibling that others cannot understand, and so on. At this point the enumeration of evidential supports for the concept is not only to clarify the concept's meaning, but also to add to the "scope" of reality to which our concepts pertain. This enlargement of the "scope" of pertinence adds "systemic import" to the concept, a Hempelian notion (p.54) that we will develop in a later section.  


Part IV: Sharing Affective States: Contagious Affectivity

This is Prof. Stern's third example of "intersubjectivity." Once again, the supportive evidence for the occurrence of "sharing affective states" comes from the research of various colleagues, such as R. M. Emdee and others.(1978, 1983) Prof. Stern correctly offers us Emdee's own operational term, "social referencing." (To give our colleagues' own conceptual terms is both scientifically elucidating and ethically correct.) The example involves the observation of one-year old infants placed in situations where there is perceptual uncertainty. For example, an attractive toy is placed at a short distance from the infant. The infant sees the toy and wishes to obtain it. But between the infant and the attractive toy there is a "visual cliff," that is, an apparent drop-off which is mildly frightening for the one-year old.  Other ambiguous stimuli may be used such as a Star War type flashing robot that may be attractive and yet slightly frightening. Will the infant approach the ambiguous stimulus or avoid it from fear? "The infant looks toward the mother to read her face for its affective content. If the mother has been instructed to show facial fear, the infant turns back from the "cliff," retreats, and perhaps becomes upset. If the mother smiles at the robot, the infant will too "and might even approach the object." (p.132) This observation conforms to the idea that the infant seeks signals that indicate the mother's affective state and immediately shares that same affect, thus determining his approach or avoidence behavior in face of the ambiguous situation. We might call this "contagious affectivity," that is, the mother's apparent (experiential) feeling becomes the infant's experiential feeling, the experience-to-experience bridge being the mother's facial expression.

The Wish for Affective Sharing
The next example, which Prof. Stern takes from research within his own laboratory, deals with the sharing of affective states from another point of view. This time it is not "contagious affectivity", but rather, the desire to discover in the other person one's own affective mood. To understand the difference, we can think of how moods are "contagious," from person to person, whether or not the moods are positive (joyful) or negative (angry or sad). This can make us think of certain emotional dynamics within a family's functioning. Meanwhile, Stern's second conceptual interpretation of "sharing affective states" is, we shall see, the desire to find in the other a representation of one's own mood. We can think of how we like to hear joyful music when we are feeling joyful, and sad music when we are sad. The observation that Stern cites is as follows: An infant is separated from the mother, and then shows behavioral evidence of still feeling sad after the mother's return. The infant is then presented with two faces, a happy one and a sad one. The infant will show preference for looking at the sad face. Conceptual interpretation: The infant is seeking to find an expressive face that duplicates its own mood.

Alternative Interpretations of the Evidence
Let us use this last example to point out that there is often a gap between the operational concept and the evidence presented. Therefore (and this is a part of the scientific adventure), more than one "conceptual interpretation" can be given for a single observation.4A To return to the above example, Prof. Stern proposes that the infant is wishing to see a face (an expression) that indicates a person with a mood (an experience) similar to his own. But another interpretation is also possible: At the moment that the infant's own face is showing an expression of distress, it is possible that the infant prefers to have a visual representation of a face that is like his own, since the tendancy to "mirror", (a person-to-person behavioral tendancy that Stern also develops in Chapter 7) means that the infant will tend to copy the expression of the face he sees. To copy a sad face when one is sad could be preferable to copying a happy face when one is sad, since the latter type of "incongruous" copying could create a tension in one's own facial expression. Thus, this interpretation would eliminate the idea of wishing to attune oneself to the other's invisible experience, and proposes, instead, that the wish is to copy an expression that is identical to one's own. The epistemological point of the above "alternative interpretation" is that connections between operational concepts and their observational manifestations cannot be taken for granted as one hundred percent certain. Just the opposite: Scientific inquiry requires a continuous reflection regarding our supposed concept-observation connections. This also means that multiple instances and multiple forms of data for a single operational concept are always welcome.

Scientific "Intersubjectivity" Encourages Greater Precision
There is another epistemological implication of Prof. Stern's treatment of "sharing affective states," which he also calls "interaffectivity." (To create a single word term for a multiple word concept is epistemologically correct, so long as there is clarity, which, it seems to me, is so in this case.) We have shown how "interaffectivity" refers to at least two different processes: 1. "Contagious" interaffectivity whereby the infant spontaneously feels the mother's apparent (facially expressed) feeling (the "visual cliff" example).  2. "A desire for" or "search for" interaffectivity whereby the infant shows preference for the sad face image in order to discover a feeling experienced by the other that matches its own feeling experience. In fact, implicit in this second category of "searching for" interaffectivity, there are two sub-categories: (a) Wishing that the Other follows one's own affect (as in the example of seeking the sad face when one is sad), and (b) Wishing to follow oneself the affect of the Other (not evidenced here, but implied by the example of looking at the mother in order to know whether the "visual cliff" or Star War robot is dangerous or playful). In fact, this latter possibility to desire to follow and feel what the Other is feeling, which will probably emerge more fully only when the infant is older and at a later stage of psychosocial development -- may be connected to the two processes of "spontaneously giving empathy" and "wishing to give empathy." The epistemological point is that the original definition of "intersubjectivity" cited from the work of Trevarthan, "a deliberately sought sharing of experiences about events and things," can be enlarged and made more precise in the following way: In intersubjectivity there are at least three types of processes:

Looking at the above theoretical construction, there may be a case (with advantages and disadvantages) for changing the wording of the fundamental concept, "intersubjectivity", and inventing a new term: "inter-experientiality". While a "subject", a person, lives an inner experience and reveals himself through behavioral expression, andthus "intersubjectivity" means the sharing of both, the term "inter-experientiality" focuses upon the sharing or coming together of both parties' experiences. Intersubjectivity has the advantage of being more inclusive. But since we already have a word for the interchange of expression and behavior, that is, the term "communication," the new term, "inter-experientiality", focuses upon what is original in this thesis, namely, that two separate processes that are actually "invisible" (unobserved and unobservable) between and among people, the realm of "experience," can be brought together and shared, even in early infancy, by means of the visible behavioral bridge, communication. Once again, the point here as throughout this examination, is to highlight the epistemological process of scientific inquiry. Thus, the last point is to show that after a review of concept-evidence examples, it is possible to have reason to modify the original overall concept, in this case, "intersubjectivity."

A Protocol For Examining "Observation-Operational Concept" Interactions
Let us make a short review of our method of elaborating the scientific adventure as evidenced up to this point. We have profited from Prof. Stern's extremely clear and documented theory construction regarding the notion of "intersubjectivity" by examining his work according to four basic methods of scientific inquiry:

Other methods, such as the important question of evaluating "systemic import", will be developed in the final part of this article, while we now continue to apply the epistemological principles of Carl Hempel to the research of Daniel Stern. The following part of Prof. Stern's scientific theory construction focuses upon a particular sub-category of the general notion of "intersubjectivity," namely, "the sharing of affect states."  


Part V: The Scientific Adventure    

Science as Drama
Prof. Stern's first cites concepts already known in this area, such as "mirroring" and "empathic responsiveness." (The reader might ask which references are the sources of these concepts.) He then restates the argument in more operational terms: "What are the acts and processes that let other people know that you are feeling something very like what they are feeling?" To make more concrete, more clear, more operational, more precise, our basic notions and questions, is a constant task within the scientific project. How many presumably "scientific works" completely neglect this crucial step and instead swamp the reader with a series of relatively abstract terms -- perhaps overlapping, perhaps coming from different levels, perhaps even contradictory -- that we are expected to swim within but not drown?   Prof. Stern then homes in on the term, "imitation," describes it more concretely, and suggests its insufficiency: "To accomplish this interaction of sharing affective states, the mother must go beyond true imitations,"(p.139) and with this criticism of a previously accepted concept, "imitation", our research ship embarks on a new adventure. (Nevertheless, the process of "imitation" is respected as an important basis for certain mother-child interactions, and Prof. Stern is unstinting with his numerous citations of colleagues' works. He gives 11 reference citations.(p.139) This not only shows that he is aware of the research accomplished by his colleagues and is not putting on a one-man show, but also offers the pertinent references and thus gives a helping hand to other researchers and students who wish to take the same investigative voyage. Once again, science is an intersubjective process of investigation). As Hempel says, "Hypotheses must be capable of a uniform intersubjective test". (p.45)   Prof. Stern then examines a particular research reference, that of Papousek (1979), to show that while there is an intra-modal (voice-to-voice) imitation-like process going on beween mother and child, the imitation is not a precise echoing of one another's sound emissions, but rather, "the mother is constantly introducing modifications or providing a theme-and-variation format with slight changes in her contribution at each dialogic turn."(p.139)   If one reads Prof. Stern's account, we can see that he accomplishes two tasks up to this point. 1. He first offers conceptual reasons for the insufficiencies of the notion of "imitation" as a means for explaining "affective attunement." (I will not review his reasoning on this point because of its complexity.)  2. He then shows, by drawing upon Papousek's observation, that even when "imitation" is observed within a mother-child research study, the "imitation" is not a total, one hundred percent to-and-fro copying, but rather, is replete with modifications and variations on a theme.   I propose that the scientific adventure is sometimes like a dramatic theater piece. A problem is presented -- "how does affective attunement occur?" -- obstacles are introduced -- "the notion of 'imitation,' already known, is insufficient -- and so we are waiting for the 'denouement' of the drama: How will the problem be solved?    

Everybody Can Play as Umpire or Judge  
It is at this point that Stern's specific contribution and solution, the notion of "affect attunement", is introduced. And then the drama unfolds to everyone's satisfaction: He produces six specific examples to illustrate operationally (that is, concretely) what affect attunement looks like and sounds like! And the audience is convinced. (I am, at least.) Another metaphor: The ball-game is won. And the spectators are glad to have bought their tickets. This is a model of good scientific progress. I can only ask the reader to read Stern's examples (p.140-1) to see if he too finds the argument convincing. And this is what science is about! An adventure and a game. Does the evidence convince? Is it adequate? Is the concept well-delineated? Does it help us understand reality? Everybody can play, because the competent scientist presents his evidence, and the reader, like a critical jury, can decide for himself if the argument is convincing or if the mystery remains unexplained.   Stern's presentation merits at least one exemplary citation:   A ten-month-old girl finally inserts correctly a piece in a jig saw puzzle. She looks toward her mother, throws her head up in the air, and with a forceful arm flap raises herself partly off the ground in a flurry of exuberance. The mother say, "YES, thatta girl." The "YES" is intoned with much stress. It has an explosive rise that echoes the girl's fling of gesture and posture.(p.141)   It is after this presentation of six clear-cut examples that Stern then presents his concept of "attunement". In other words, his definition arises as an induction, that is, a conceptual interpretation that is given after the specific phenomena that illustrate the domain of inquiry are presented. Of course it would in no way be incorrect to present ideas and evidence in the reverse order, starting with the definition and characteristics of the key concept and then presenting the supporting evidence. However, to take this opposite tack and present first evidence and then the concept invites the reader (in my opinion) to reflect upon the entire matter in an active way: Is the evidence sufficient? Do we have a sense of "what really happens"? Is the author's abstraction of a common element in these examples the same as my own abstraction? Do we understand the same thing when we are faced with the same events? Can we play umpire, jury or referee and decide "who wins" or "who done it" because the evidence is presented before our very eyes and we have refused to be duped by the clamor of prejudice?   In fact, whether the researcher presents his evidence first or else presents his major concepts and hypotheses first is a secondary matter. The main point is that the critical reader or colleague is offered the opportunity to join in the game and make an independent judgment: Are the researcher's conceptual interpretations "correct" or, better, "adequate", in terms of corresponding to the facts, and do the concepts cover the evidence in a sufficient way? Prof. Stern's material, in my opinion, fulfills in a satisfactory (and, I propose, sometimes remarkable) way all of these conditions, when he elaborates his theoretical-empirical research regarding "affect attunement".   What am I trying to point out here? The epistemological thrust of Prof. Stern's work is that science is the obtaining and presentation of sufficiently numerous and detailed observations, all put into interplay with their explanatory concepts. In a word, science is the interaction between observations and concepts. This can modify the traditional (and discouraging) outlook of science, as already mentioned, whereby the professional worker (scientist or clinician) believes that social science must adhere to the physical science paradigm whereby we are obligated to create experimental groups with variable "X" and compare them to matched control groups without variable "X" in order to produce adequate scientific data. Stern's work belies this prejudice. What is Prof. Stern's data regarding "affect attunement", up to this point, in his presentation? Six descriptions of mother-child interactions. And from this empirically derived data he presents his conceptual explanations. This is science. The data is sufficient in number and precision, so that the intelligent reader can reflect, for himself, whether the concepts presented to explain the data are "adequate" (not whether they are "true", a criteria used for religious faith and support, but not for scientific observation and hypothesis-making).    

Hubert Montagner  
We might, at this point, cite another scientific work in developmental psychology -- that of Prof. Hubert Montagner of the University of Besancon -- who made a film of children's behavior in a nursery school setting. The theme regards child personality development: "Is the child social? a leader too? or else either withdrawn or inappropriately aggressive?" These questions are shown to be connected to the child's capacity or incapacity for mirroring behavior, and this, in turn, was related to mother-child tendancies toward reciprocal mirroring. Once again, even without statistics regarding experimental and control groups, even without, in fact, any quantitative measures, but merely with the filmed presentation of a single situation that revealed a sufficient number of relevant examples, the researcher in psychology (developmental or clinical), can engage in the scientific project of recording and presenting empirical evidence and, following this, plunge into the adventurous domain of scientific theory construction.    

Cross-Modal Attunement  
On the conceptual level, Prof. Stern proposes that "attunement" interactions show a correspondence, between the two parties, at the level of "intensity, timing and shape."(p.146) These three categories are then sub-divided into six categories that are yet more precise. (At this point, the evidence and conceptual considerations that Stern presents are too detailed for our purposes. The reader is once again encouraged to go to his original work.) And the consequence? A new conceptual refinement, "cross-modal attunement", is then presented. This concept can explain (Stern proposes, and I agree) the observed fact that when a child makes one type of expression, that is, in one modality, the parent might respond within another modality of expression. The example cited above shows a gesture-voice cross-modal interaction. (Little girl's arms flap up, mother says, "Yes! Thatta girl.")   As the reader works his way through Chapter 7 on "Affect Attunement," he can only be struck by the quantity of examples that Stern constantly brings to light in order to support, and at the same time clarify, the key concepts. If we take this as an epistemological model, what does it mean? It means that the researcher must do a great deal of work, taking from his own and from colleagues' research studies the specific examples that illustrate and support his basic conceptual themes. And everything drawn from colleagues is cited with the exact references, which is only a matter of respect as well as "pointing the finger" to draw the reader's attention to "the target."        
Up to this point, our main argument was to examine Prof. Stern's two chapters on "intersubjectivity" and "affect attunement" in terms of the epistemological question, "Is this work empirically sound?": Are we clear as to the nature of our observations and operational concepts? What is their meaning? Can we break down their meaning still further into sub-categories? Are the two realms connected? Is there sufficient amount of evidence? Is the hierarchy of concepts -- from the overall concepts to their operational components -- well-delineated?   The last epistemological question that we shall ask -- always profiting from Stern's exceptional work -- is whether his study is important? Is it fruitful? Is it relevant? And this brings up, as we shall see, another series of considerations that are essential for scientific theory building.   However, before changing our navigation chart in order to enter into such new waters, let us turn back the clock and compass to cite what Carl Hempel has to say about empirical precision and clarity.    

Determinacy and Uniformity of Usage  
Hempel says that we must be very rigorous, in two senses, when we consider the clarity of our scientific language (evidence or concept):   The meaning of a given expression presupposes that the conditions of its application are (1) well determined for every user of the language and are (2) the same for all users. We shall refer to these two presuppositions as the conditions of determinacy and of (personal and interpersonal) uniformity of usage.(p.10)   In more colloquial terms, "Do we know what we mean when we say what we mean?" Do we know what we're really talking about? Are our ideas clear? (the concepts) Are our references to reality clear? Thus, when we did our work of "analytic reduction" in Part I, we were clarifying our ideas. And when we were examining our observation-concept relations, that is, our "rules of correspondence," we were checking up on our references to reality. In this sense, we were examining the determinacy of our terms.   Hempel's second point -- that our ideas and observations must be the same for all users -- can only be confirmed when colleagues meet together and discuss the scientific work. This is the "intersubjectivity" criteria of science; science is also (but not exclusively) an interpersonal process which presupposes, at some point, a direct interaction among various workers in order to examine the acceptability, presumed validity, or supportability of its conceptual and evidential features. My guess is that these two chapters of Prof. Stern's scientific project would pass this phase of examination with flying colors. But that is only a guess. Science requires that the interpersonal event actually occurs through congresses, group presentations, seminars, etc.4B  

Reliability of Our Observations  
Although Carl Hempel's treatise emphasizes the determinacy and uniformity of usage of our conceptual terms, the same criteria of rigor apply to our observational base: the data we collect. In other words, as we collect our data, is the meaning of our terms used for our observations sufficiently clear (determinacy), and would the same observer looking at the same phenomenon at different moments give a constant result? (intrapersonal uniformity) And, so important in all science, would different observers looking at the same phenomenon give uniform observations? (interpersonal uniformity) Let us call the combination of all these factors the "reliability" of our data.   Now let us see the particular challenge that these criteria of "reliability" raise in the social sciences: psychology, anthropology, sociology, etc. In the physical sciences, our means of obtaining reliable data is usually through the use of some obvious physical measurement. For example, we may have an electrical apparatus with a pointer that indicates from 1 to 100, and the observer's eye, or, more frequently today, another machine, notes down the movement of the pointer. Therefore, the capacity of obtaining reliable data, in terms of repeated registration of the same observation, is one of the strongpoints of the physical sciences. And, I propose, this represents one of the weaknesses of the social sciences. For when we try to observe the event we are trying to understand -- human beings in interaction -- different observers will "see" different phenomena and come to different conclusions. In fact, in order to obviate this fundamental difficulty of obtaining reliable data, the tradition of experimental psychology has been to create "the laboratory experiment": By restricting the observed situation to one that takes place in the laboratory, and then, by restricting yet further the phenomenon to be observed to a special task that is never found in ordinary human life, such as certain types of learning tasks or perceptual tasks, or, still even more reductive and further from human life as it really exists, by changing the subject of our studies from human beings to rats, in order that the data be "reliable" (and, in this sense, "scientific"), the experimental psychologist, yes, has faced the epistemological question of "reliability." But at what cost? (We shall discuss the limitative consequences of obtaining such "reliability" in the next section, when we ask ourselves, "But is it important?")   And what is our proposed solution?    

Use of the Video Camera Represents a Giant Step Forward in Social Science  
"Technological advance can create scientific advance." Daniel Stern and his host of colleagues in developmental psychology now use the video camera in order to register mother-child interactions. This has helped psychology as a science take a giant step forward. The highly complex event of interpersonal relations can now, finally, be registered. And even though the registration usually occurs in the laboratory, the instructions given to the mother, "Can you play with your child as if you were at home?", within a physical setting of tables, chairs, play objects, carpeted floor, and so on, that reproduce the natural setting at home, and within a warm and empathic rapport that decreases the sense of being judged or evaluated, all this, permit a certain degree of similarity between the laboratory and the natural field. Is this "certain degree" of similarity between laboratory and home sufficient? I propose that we do not yet have the evidence to decide upon this. But the reproduction of the home situation within the laboratory and its then being recorded by a video camera is certainly a large step, perhaps even a giant step, with respect to earlier protocols for laboratory investigations in psychology.   How can the thesis -- that the laboratory reproduction of the home situation reveals processes "sufficiently close" to the home reality that it merits the researcher's efforts -- be verified? One answer is as follows: Cameras are now being put into homes and are running constantly. So that future research will indicate if the two situations, home and laboratory, turn out to be sufficiently similar in order that laboratory studies remain justified.5      

Evaluating the Filmed Observations  
The problem of obtaining "reliable" data collection -- to follow the rigor of Hempel's epistemological criteria of determinacy and uniformity --takes place in the second step of data collection: the evaluation of the film sequence produced by the video camera. As different observers see the same film strip, will they see the same thing? And then, will they note down the same observations in their data note-book? For the data in science is not the film itself. The film becomes the "phenomenon", now registered and so available for repeated observations, and the written down observers' evaluations become the new raw data that will be tabulated, synthesized, sometimes transformed into graphs, then published and discussed. And here is the epistemological barrier when we use video camera registration: Will different observers give sufficiently similar descriptions, make sufficiently similar quantitative and qualitative judgments, of the video filmed phenomena, so that the collected evidence is reliable?   The problem is noteworthy, but not impossible to overcome. 6
The data produced by Prof. Stern and the data that he cites from his colleagues' works seem to have sufficiently overcome this barrier, though I imagine that their research groups are making constant revisions and refinements of their video tape evaluations in order that reliability becomes increasingly certain.        


Part VI: Scientific Importance     Sacrificing Importance for Convenient Observation  

As we said, the empirical question of scientific theory construction must be followed by another fundamental demand: "Is this scientific work of any importance?" How does Carl Hempel help us map out this new epistemological terrain?   He starts out with a joke: We might define the 'hage' of a person as the product of his 'height' ('h') in millimeters and his 'age' in years. (Thus, 'hage'.) This definition is operationally adequatae and the term 'hage' thus introduced would have relatively high precision and uniformity of usage. But it lacks theoretical import, for we have no general laws connecting the 'hage' of a person with other characteristics.(p.46)   In this imaginary situation, Prof. Hempel shows us the degree of absurdity we may reach if we sacrifice pertinence for scientific precision. This is like the story of the man who is found, late one night, looking under the street lamp for his lost wallet. "What are you doing?", asks his friend. "I'm looking for my wallet." "What happened?" "I lost it over there!", pointing to some dark shrubbery in a near-by garden. "So why are you looking for it here?", asks the concerned friend. "Because this is where the light is!"   Is this what science is? Looking in the wrong place because that's where we can so easily point our investigave light? In a previous section I mentioned the "cost" of certain types of reductive laboratory studies in experimental psychology because that's what we can see, quantitatively measure, write it down, publish and call it "science." Yes, it's precise, but it's not "where the money is" in terms of scientific profit, fruitfulness and productivity.   Perhaps we scientists should act a little bit more like Willy Sutton, the famous bank-robber of the 1950's: "Willy, why do you keep robbing banks?", asked after his sixth arrest. "Because that's where the money is!"   So where is the profit in science? How can such difficult notions as scientific "productivity," "fruitfulness," "importance", etc., be clarified? Hempel tackles this job, and we will follow his lead:    

The Study of "Natural Systems" And The Weakness of Questionnaires  
Hempel proposes that the goal of science is to study and explain "natural systems."(p.53) That means that the social sciences are aimed at studying human reality: psychological, social, anthropological, economic, political, historical, etc. And the "natural system" of human reality is not what we find in the restricted laboratory setting, but that which occurs in the "natural setting" or, as in developmental psychology, in the "open" laboratory setting that attempts to and succeeds in reproducing, as much as possible, "real" human reality, human reality as it occurs spontaneously in non-scientifically obsereved life. Perhaps, as in quantum physics, where we can never measure precisely the electron's position and motion at the very same instant, or, as in relativity theory, where we can never discount the observer's influence, it may be that we may never achieve the absolute scientific study of non-scientifically observed realities. But let's not throw in the towel. We can come close. The technological advance of the video camera has given us high precision regarding the visual and auditory registration of expressive behavior. (It is not sure if "invisible experience" will ever be externally observed and recorded on a direct basis, but here we must admit our limits.)  
This rigorous definition of science, namely, as the study of "natural systems," such as human reality as it spontaneously occurs, not only highlights the methodological insufficiencies of the scientist who uses the restrictive laboratory system (as already differentiated from the "open" and somewhat "natural" laboratory system, as in today's developomental research), but also points to the methodological inadequacy of one of social sciences' favorite tools: the questionnaire! We have real occurences in daily life among people, "the natural system," and then we ask people to write down their answers to a series of pre-established questions as a manifestation of that reality, as "evidence" and "facts" for a presumed scientific study. And even if the questions are "open-ended," we are asking for a verbal recounting, in memory, of a complex reality, especially when it is interpersonal and involves an awareness of non-verbal interactions where there are essential two-person (and even multi-personal) variations of rhythm, intensity and shape, as Prof. Stern has proven so conclusively to be of essential determinative relevance. Thus, I propose that when using such methodologies, the distance is too great! At least for when the field of study is a complex and sequential human interaction, and not a simple behavior that can be defined by simply remembering "what did I do (like watching television), how many hours", etc. And even such simple behaviors will often be distorted when registered by the questionnaire that attempts to record, according to fragile human verbally-defined memory, "What actually happened." 7  
A recent methodological innovation, called "the Experience Sampling Method" shows an interesting attempt to overcome the distortion of memory in registering daily-life events: The person carries an instrument that sounds off at random moments in daily life. Upon hearing the sound, the person must verbally record exactly what was happening -- what action, thought and feeling was occuring -- at the very moment that the buzzer made the noise.(Massimi, 1987) I think we can appreciate such methodological innovations that attempt to register the terrain of the "real" and diminish this distance between "real" and "registration."    

By What Criteria Can We Judge "Scientific Fruitfulness"?  
But thus far we have only treated the relatively simple question, epistemologically speaking, of adequately registering data from a situation that is as close as possible to the non-investigated, ordinary and spontaneous reality of human interaction.   The more challenging question for the researcher in psychology and for all social scientists is as follows: How can we know if our scientific project, with the particular operational and higher order concepts that we have chosen, and with the particular categories of evidence that we have selected for our attention and note-book descriptions, are scientifically "fruitful" and "productive"? For Prof. Stern, for example, how can he justify that his study of "intersubjevity" and "affect attunement," with its concepts and evidence, has scientific importance? And how can more critical colleagues debate its importance?      

Locate the "Natural System"  
Carl Hempel points to the epistemological solution: Our classifications of concept and supporting evidence are fruitful to the extent that they describe, explain and predict the phenomena of our "natural systems."(p.53) In other words, science, as Hempel says in his very first sentence, is to describe, explain and predict phenomena in the world of experience, "natural systems," which we will call, to be brief, "reality."   But our question is not yet sufficiently responded to. For by what criterica can we determine whether a presumed scientific work describes and explains reality? How can we operationalise this abstract idea?   Yes, the correlation between our observations and the concepts which explain them, is the key to science, in the sense of rooting our ideas in some realm of reality. And to this end we have the concepts of "high determinacy" and "uniformity of usage."(p.47) But one key is not sufficient to open the door toward scientific fruitfulness, as the analogy of the man looking for the wallet he lost in the dark garden by searching in the street illuminated by teh lamp, clearly shows us. Our work can lack "import."(p.46) Thus, what other scientific keys must we carry on the key chain of our scientific project? To clarify, how can we know that the reality to which our data refers, and the concepts we then generate, are part of a "fruitful" scientific project? Hempel puts forward the question in this way: How can we be sure that our scientific project is dealing with "a natural system" (p.53) meaning a relevant observational base, and how can we know that we have constructed "a natural classification"(p.47) that is, a pertinent theoretical construction ?   

Criteria of "Scientific Fruitfulness"  
One way is to show that a basic scientific format "exhibits clusters of empirically correlated features"(p.47) This means that the initial format leads to further observational work within one's own project, like, within music, variations on a theme. A second way is to show that our theorizing "enables us to make the maximum number of prophecies and deductions."(p.53) A third way is to show that the results of the project, both observationally and conceptually, enter into "fruitful connection"(p.49) with the scientific projects of colleagues who are make investigations and developing theoretical construtions within the same area and similar areas. And a fourth way is to show that the project has "scope"(p.46), that is, that it connects to other fields of interest.

New Empirical Studies, New Hypotheses, Interlacing with Colleagues and the Scope of Implications  
Let us now continue our scientific adventure of examining Prof. Stern's elaboration of "attunement", in light of these four parametere, in order to see whether he successfully fulfills Hempel's requirements of "scientific productivity." The following is a re-statement of Carl Hempel's requirements for fruitful theory construction:   1. New empirical studies that shed further light on the fundamental concepts (studies performed by one's own research team as well as cited from the works of colleagues).   2. New hypotheses generated within one's own theoretical structure.   3. The interlacing of one's own concepts with those presented by colleagues: similarities, differences, overlapping.   4. The implications for associated fields of research (e.g., personality development, parental learning, etc.) and for even more distant fields of knoweldge (evolution, art and philosophy), also known as "the scope".     And, to put the cart before the horse, we can announce the results of this examination at the very outset: Prof. Stern's research passes the test with flying colors. His work is "fruitful" in terms of all of Carl Hempel's epistemological criteria. That is why, in fact, we are presenting in such detail the work of Daniel Stern: In order to elaborate a model of scientific inquiry and theory construction that other psychologists might well profit from as they elaborate their own scientific protocol.  
We will now see how Prof. Stern's text satisfies each of the above criteria of scientific "fruitfulness":   I. New empirical studies:   Prof. Stern's research team elaborates further studies regarding "attunement," for example, by compiling data that indicate the quantitative aspects of attunement in relation to two other types of comportments, namely, the mother's making verbal comments and her showing imitation gestures at the moment the infant shows a specific affect. Thus, the researchers note that in response to the infant's expression of affect, the mothers respond with attunement behavior 48% of the time, with verbal comments 33% of the time and with imitations 19% of the time. They add that the attunements occur at about the rate of one every sixty-five seconds. This demonstrates that not only does attunement occurs, but also that it occurs frequently, meaning that it has quantitative significance; in this study, it is the most frequent maternal response to a child's affect. Thus, attunement is significant in quantity and not only in quality.(p.147)  
Another quantitative finding has to do with the frequency of "cross-modal attunement," meaning mother-child expressions that use different modalities: sound, facial expression, gestural expression with the arms, postural change, breath accentuation, total body movement, etc. "Cross-modal attunement" stands in contrast to "intra-modal attunement" where the mother-child interactions takes place within a single modality. And the conclusion? The "cross-modal attunements" were more frequent, occurring with or without accompanying intra-modal attunements 87% of the time.(p.148) Stern shows at a later point that "cross-modal attunement" has important implications for other branches of knowledge -- not only scientific -- but also artistic and philosophic.   The various parameters of attunement behavior are then quantitatively measured: Intensity matching is the most frequent, followed by matching time and then matching shape. (Prof. Stern gives the percentages. p.148)    

Are the Mothers Aware of Their Attunement Behaviors?  
Variations on the theme continue; new questions are asked: Is the mother aware of what she is doing when responding to her child's affect with attunement behavior? More specifically, what is within the mother's awareness (the experiential question) as she gives attunement to her child? How can this invisibile experience become manifested? We have already indicated the weakness of relying upon pure memory. The correct use of the registered video film offers an important solution to this methodological quandary: Prof. Stern replays the videotaped sequences to a group of ten mothers. Most mothers note the following experience: They have the wish to "share" and "be with" their children. Thus they are aware of the goal. They are not aware of their method of giving "affect attunement" by partial matching of intensity, rhythm and shape of expression. In addition, the mothers are not aware of giving matching intra-modal and cross-modal expressions.  
And when the mother is "misattuned"? What is going on within the mother at this point? Two possibilities are described (without, in this case, quantification of the data): Either the mother purposefully misattunes (stronger or weaker than the child's expression) in order to alter that expression, or else the mother does not misattune on purpose but, rather, she is out of touch with the child's expression and thereby misses the mark.  
And when the mother is in correct attunement, is she aware of this? 32% of the time "entirely," 43% of the time "partially," and 24% of the time "not at all."(p.149) But even when "aware", the mother's awareness is usually only that of the purpose of her response to the infant -- namely, "to share" and "to be with" the child during his emotional state -- while she is not aware of her attunement method, that is, of matching the child's intensity, rhythm and shape. This shows that a great deal of attunement behavior goes on outside of the realm of consciousness. (The criteria for these distinctions, that is, specifically, what questions were asked of the mother, are not offered here, but one expects that such information is clearly spelled out in the research paradigm. A competent scientist has to judge what details of a research study would be excessive and tedious and what are essential and highly pertinent for his audience. Otherwise, we run into the same problem of showing a blue-print of the house plumbing and electricity when a friend asks us where the bathroom is.)

Does Attunement 'Prolong' Spontaneous Behavior?
Stern then gives other more detailed observations regarding maternal attunement behavior: First, this behavior by the mother does not necessarily create an evident change in the infant; the infant merely goes on doing what it has been doing. (One might hypothesize that an infant, when receiving a mother's attunement behavior, prolongs his behavior over time, in contrast to the infant which is alone or with a mother who is not actively responding. To support this hypothesis would require, of course, the filming of additional sequences, and the compiling of other quantitative data. Nevertheless, it is easier for a scientist to suggest to another scientist what he should do. than take on himself a particularly arduous and time-consuming task.)    
The 'Still-Face' Experiment   Productivity in science also comes from integrating colleagues' works, as seen already in Stern's section on "intersubjectivity." In this case he cites the work of Tronick (1978) who uses the "still face" research paradigm: The parent shows, all of a sudden, a "still face" passive and expressionless -- during an on-going interaction with the infant. The infant not only stops its ongoing behavior, but also shows some degree of emotional distress and withdrawal. There may also be attempts, on the part of the infant, to re-engage its "still-faced" parent in an active interchange. Thus, we observe certain consequences of "misattunement."    

Does Misattunement Cause Inhibition and Mistrust?
Prof. Stern then offers another set of observations, this time from his own research program, and its theme is the impact of the opposite of "attunement," namely, "misattunement." This is correct logic within the scientific adventure. If variable "X", "attunement", is observed to create certain consequences, positive ones, then it seems obvious that the opposite of "X", that is, "non-X," "misattunement", will produce opposite consequences. But, for good science, "what seems obvious", from a logical point of view, is not taken for granted. The scientist, whenever possible, tries to observe the situation of "non-X," in this case, "misattunement," to see if his prediction of "opposite consequences" is born out.  
How how does Stern go about studying "non-X", "misattunement"? In the first place he asks a mother to attune herself to her child at a moment that the child is spontaneously active: Specifically, the mother is asked to jiggle the child from side to side on his bottom as he is crawling forward, creating the jiggling movement in attunement with the child's movement. To create this attunement the mother's jiggling movement matches the speed and intensity of the child's arm movements and vocalisations. Result? The child goes on crawling.(p.150) (Prof. Stern proposes that the child's response to this attuned jiggling is "nothing; that is, the child simply continues his movement without missing a beat." Another possibility, as already mentioned, is that the child's movement is prolonged in time. Why reiterate this point? Because if attunement is present, our hypothesis is that it contributes to the child's general development in "impulse formation," that is, in the initiation and sustaining of spontaneous movement and perhaps other activities.)  
And the creation of misattunement? Prof. Stern checks out misattunement in both of its directions: insufficient intensity and excessive intensity. In other words, the mother's jiggle, in one case, is more slow (the rhythm factor) and less energetic (the intensity factor) than the baby's movements. In the other case, the opposite. In both cases the infant stops crawling (impulse inhibition?) and also looks at the mother with an inquisitive expression (rupture of rapport?). Stern then adds another example of a mother instructed to "undershoot" and "over-shoot" her attuning responses within the same parameters: timing, intensity and shape. In this example, as well as in examples observed by colleagues, the same phenomenon was noted; behavior interruption (inhibition) and an inquisitive look as the child looks at the mother (a slight disturbance in the on-going rapport).8      


Part VII: Enlarging the 'Scope' of the Consequences     Personality Theory And Behavioural Reinforcement Theory  

Such data, such experiments may have "important theoretical implications and consequences", and thus, as Hempel says, can buttress the "scientific import" or "fruitfulness" of the research project. Let's briefly look at one interesting implication: Is it possible, for example, that such misattunements represent an important dynamic in personality disturbance? 8A Could misattunements between mother and child, when they are too frequent and too intense, account for the inhibition of the child's impulses and for a disturbance in the mother-child relationship, leading to personality difficulties later in life such as chronic inhibition and constant lack of trust? Thus, the scientific relevance and fruitfulness of Stern's project, becomes increasingly supported. Based merely upon this particular observation, his work can have significant consequences for understanding normal and abnormal personality development.     Another aspect of "fruitful scientific theory construction" is its interlacing with other known theories. Stern mentions that the positive results of attunement might be explained, in part, by the well-known behaviorist theory of "positive reinforcement." He points out, nevertheless, an essential difference: Attunement requires "matching intensity" while "positive reinforcement" speaks only of an "absolute level" of intensity. Thus, attunement fits the data more adequately than "positive reinforcement." But Stern does not discount in an absolute way that positive reinforcement can take place: "There is no problem with attunements also serving reinforcing functions. But simple reinforcement cannot explain away attunement. The two phenomena are undoubtedly embedded one within the other."(p.151) I quote this point to show how Stern refuses to take a dogmatic position. The dialogue is still open.    

Cross-Modal Attunement  
Another consequence of this study is the implication of cross-modal attunement. First, Stern cites other investigators' evidence that this phenomenon does occur, and even frequently. In fact, it is worthwhile to present the fascinating work of Meltzoff and Borton (1979): Three-week old infants are first blind-folded and then given one of two different pacifiers to suck on. One pacifier has a smooth, spherical-shaped nipple and the other has a nipple with nubs protruding from various points on its surface. After the baby has experienced touching the nipple a number of times, only with the mouth, the nipple is removed and placed side by side with other kinds of nipples, and the blind-fold is then removed. What does the baby predominantly look at? The nipple that it has just been sucking! Whether that be the smooth and round nipple, or else the nipple with the nubs! So there is a transfer of the sensory pattern from the tactile feeling of the mouth and lips to the visual field.

Cross-modal attunement.  
How is it possible to explain this universal phenomenon? Stern proposes that the brain must, in some way, create "abstract representations" of sensory patterns that then permit linkage among different modalities. He doesn't, in this work, suggest more about the possible underlying neurological structure that can perform this operation of going from concrete-to-abstract-and-abstract-to-concrete. But he does give examples that illustrate this to-and-fro movement between sensory modalities and their abstract representation, drawing upon other researchers' observational studies and theoretical constructions, as well as various scientific, philosophic and artistic works throughout history, going from Aristotle to Baudelaire to Susan Langer to Jean Piaget to Charles Darwin, and so on.(pp. 48-68) This, to my mind, is scientific "fruitfulness", according to the Hempelian parameter of "scope", developed to an ultimate degree.  
And, in a final section, Prof. Stern suggests that the attunement process will be the stepping stone toward other key developmental phases such as the development of language.  
The reader, too, can add "scope" to a researcher's theoretical elaborations by means of his own assocative process. Let me offer two examples:   I. Parental Training: The concept of "affect attunement" as a fundamental modality that creates both a positive rapport between mother and child and a reinforcement of the child's spontaneous behavior, suggests that parents may well be instructed -- by conceptual lessons as well as practice sessions -- to create "affect attunement" and thereby avoid the mishaps of chronic misattunement. The fact that mothers were not aware of this universally present positive interaction, whether or not they were applying it from an intuitive sense -- and, more striking, the fact that even child development researchers were not aware of its present before the use of video camera filmed studies -- implies that merely to bring parents' attention to the fact of this intersubjective rapport can represent an enlightening discovery.   II. Psychotherapy, Personal Development and Education: The Biosystemic School (Liss and Stupiggia, 1995) and the Biosynthesis School (Boadella,1987) of Body Oriented Psychotherapy, the School of Hakomi Therapy, (Kurtz, 1983) as well as other approaches toward personal development such as Neurolinguistic Programming, (Bandler, 1979) had been using "partial mirroring", "correspondence of gestures," and related "attunement-like" methods even before Daniel Stern's studies were known. Nevertheless, his exposition reinforces the application of "affect attunement" to the therapeutic session. Going further, the consciousness of the parameters of cross-modal attunement invites the professional therapist and the group leader of creativity groups to become still more inventive and refined in the elaboration of "attunement" in both verbal and non-verbal behavior, especially since it appears to catalyze the three major principles of personal evolution: spontaneous and authentic expression, positive self-esteem and a trusting relationship. Even teachers' training courses are now profiting from this fruitful awareness and methodology. 9    


SUMMARY:  A Two-Step Protocol For Scientific Observation And Theory Building In The Social Sciences  

What have we done in this work? We have presented ceertain key concepts offered by the philosopher of science, Carl Hempel, and illustrated their application to psychology by taking a research study of unusually high quality, that of Daniel Stern, as presented in Chapters 6 and 7 of his book, The Interpersonal World of the Child. The purpose is to give specific indications for researchers in psychology, not only for those in developmental psychology, the specific domain of Prof. Stern, but also for those in clinical psychology, where the researcher is also the clinician who is ready to use his unique case studies as the observational baseline for scientific theory formation, perhaps even taking his data, as Prof. Stern does, from video camera studies of the therapeutic session, that is, from the natural system.  
Here is a review of Carl Hempel's epistemological format that the clinician in psychology (as well as other social scientist researchers and practitioners) might consider following:  

Definition: Empirical science has two major objectives: to describe particular penomena in the world of our experience and to establish general principles by which they can be explained and predicted.  In a sense, we are in a two-step process:
Step I. Link our ideas to observations. II. Order our ideas and generate new ideas, in interaction with the observations and ideas of others, so that we can consider our work "relevant" and "fruitful."  
Step I: Link our ideas to observations: 1. Reduce our general concepts to more precise ideas that almost touch reality. This means to define our operational concepts. 2. Define the specific observations which we wish to make which correspond to our observatinal concepts. 3. Register observations from the "natural reality" we wish to study, or from the laboratory reality which attempts to approximate the "natural reality." Thus, for the clinician, the registered "natural reality" that we wish to study is the therapeutic encounter. 4. Create qualitative verbal descriptions and, when possible, quantitative compilations, using several observers when possible, of the reality we have registered. 5. Create a written presentation that offers concrete examples of our observations and of the quantitative data, when available, all presented in conjunction with the operational concepts that we are seeking to support. 6. Relate these operational concepts to higher order concepts and present the structure of the theoretical matrix.    
Step II: Show that our work is relevant and fruitful: 1. Show new empirical studies that come from one's own work, as well as the from the work of colleagues, that contain observations that further support and deepen our conceptual explanations. 2. Indicate the new hypotheses and refinements of earlier hypotheses that emerge from these additional observational studies. 3. Interlace our operational concepts and higher order theoretical constructions with those presented by colleagues, indicating similarities, differences and overlaps. 4. Spell out the implications of the research observations and theoretical constructions for wider fields of study (the "scope") that are within our own branch of psychology as well as in other fields of science, as well as, at times, in other domains of human endeavor such as art and literature.   Conclusion: The scientific adventure lays before us. This presentation is to offer a map for the navigator, who wishes to explore new waters, and also for the passenger, who wishes to follow the route with his own sensitivity and decide for himself, "When this captain steers me through unknown waters, do I agree with his interpretation of the route and the importance of the destination?" Or is he merely sending up a balloon that would be best to puncture and deflate without wasting more time on the matter?

Cited from "The Philosophy of Science and Clinical Research," published in italian as "La Filosofia della Scienza e la Ricerca Clinica," Psicologia Clinica, (Ed. Prof. M. Bertini), No. 2,, 1983, and in La Psicoterapia del Corpo, (by Boadella, D. and Liss, J.), Rome, Ed. Astrolabio, 1986.   While in the past our field of study required us to make repeated observations involving situations with one or two variables, our field of study today (within the social sciences) involves "complex and unique events." This becomes the new observational base of our science. Thus, the scientist's goal changes target: Rather than register and compile simple observations that involve the independent and dependent variables (input and output) in a series of repeated, experimentally controlled situations, the clinical researcher, by turning toward the principles of General Systems' Theory, is encouraged to make "complex" observations within "a unique and unrepeatable" situation. Thus, his first task becomes that of collecting complex and pertinent data.(p.158)    

2. (op. cit.)   In traditional science the data is precise and simple. The "effort" of science consists of the painstaking task of creating experiments with the goal of obtaining reliable data.   But the "effort" of science in the social sciences changes direction. At the moment of obtaining data, the "effort" of the social scientist is channelled toward accurately registering all of the pertinent observations. This requires less effort over time, since it is not necessary to keep repeating the same experiment over and over as in traditional science. On the other hand, this approach requires a greater effort during the period of observing the registered data and compiling the results, which means either giving a description or else a quantitative analysis. Thus, to observe complex and non-repeatable events and then transmit the observations into a transmittable form requires new capacities on the part of the scientist.(p.160-161)   2A. One might ask why I have chosen a work in developmental psychology while the thrust of the argument is that clinical psychology needs a more fundamental epistemological format in order to become scientifically grounded. The reason is that I have not found any work in clinical psychology which offers the precision and completeness of Prof. Stern's research study and presentation. But, it may be objected to, the observational methods of developomental psychology are not those of clinical psychology. But that is no longer the case. Video taped observations and evaluations are the basis of developmental psychology today. This is the point! The motivated clinical researcher now has the opportunity to use video-recorded registrations of his clinical work as the basis of his scientific project. More will be said on this crucial point in a later section.  

3. Laing writes in The Politics of Experience,   I see you, and you see me. I experience you, and you experience me. I see your behavior. You see my behavior. But I do not and never have and never will see your experience of me. Just as you cannot "see" my experience of you...   I cannot experience your experience. You cannot experience my experience. We are both invisible men.(p.17-18)  

4. See Liss, Jerome, The Meta-Communication: Talking About Talk in Order to Talk Better," (1985, unpublished paper) for an exposition of the communication problems engendered by failing to recognize this distinction between visible behavior and invisible experience, as well as for the presentation of a rigorous form of communication, called "the meta-communication," (though different from Paul Watzlawick's concept of "meta-communication"), that aims to overcome this problem.  

4A. "Conceptual interpretation" refers to the "ascending movement", to return to Hempel's spacial analogy, among "strings" that connect the lower plane observational "knots" or "nodal points" to the higher plane operational concepts. We can call the "descending movement" from concepts to observations, "observational grounding." For the total back-and-forth relationship, Hempel offers us two terms: "rules of correspondence" (citing Morgenau) and "epistemic correlations" (citing Northrop). (from Hempel, p.82)  

4B. Can "group discussions" through E-Mail also offer us a new technological force to permit interpersonal contact among colleagues? I believe so.  

5. In any case, the use of the video camera will be widely extended: The camera will be put in the home more and more frequently, where it will be kept running all day long; people will get used to it and then forget its presence, at least in regard to ordinary comportments (obviously, not sexual, nor those involving extreme emotions, nor the discussion of money). In addition, since we are at the frontier of important changes, because of video camera registration, in the field of human psychology, it is important that the researcher - clinician clarify the epistemological status of his concepts and evidence.  

6. I have examined this question in some detail for the following reason. Prof. Stern and his colleagues in developmental psychology are moving ahead in a rapid and systematic fashion in order to resolve this problem of using video filmed sequences as a means to create a data bank replete with reliable evidence. We must remark -- and this is not an insignificant point -- that these scientists are usually research psychologists paid by a University and sometimes receiving research grants. Therefore they are using the financial and interpersonal resources inherent in their situation as a means of facing and overcoming the epistemological problem, that of creating reliable data of interpersonal interactions based on the observation of video film sequences. When we confront the problem of how clinical psychologists can use film strips to make a scientific study of their clinical sessions -- the subject of future articles within this epistemological project -- we will observe that the problem of interpersonal reliability becomes a formidable obstacle, not only because of the difficulties that the developmental psychologist must also face, but also because of the lack of material and interpersonal resources that characterizes the situation of the clinician who works alone.  

7. To study the therapeutic process, for example, by means of questionnaires, would also be totally insufficient for the same reason of too much distance between the two realities, namely, between the "natural system" being studied, the therapeutic encounter, and the registration of what happens by means of a questionnaire. Relying on therapist notes written after a session is also methodologicaly weak because of the same factor of "distance." Therapists who can write down everything said by themselves and their patients during the therapeutic session are, admittedly, trying to overcome this "distance." But how does the client feel about this? And does not the note-taking itself interfere with the "natural" therapeutic process in which no notes are taken? If a colleague or student writes down everything he sees or hears during the therapy session, especially if this is done behind a one-way screen, as often occurs in the family therapy setting, we have much more proximity between the "natural event" and the "registered event." But, then, at that point, why not use the video camera? And then take notes of the video sequence? (Such questions will be elaborated upon in greater detail in subsequent epistemological articles that will focus upon the scientific problem of studying the psychotherapeutic process.)

8. For a more thorough discussion of the interaction among emotions, interpersonal rapport and self-esteem, see "The Self, the Impulse and the Other: Three Models of Psychoanalysis," (J.Liss) in Energy and Character, (Editor: David Boadella), Vol. 23, No.2 and Vol. 24, No. 1, Sept.,1992 and April, 1993.(pp. 21-32) 

8A. The fact of creating, artificially, a disturbing situation for an infant or for any person for the sake of scientific discovery may raise certain questions of ethics. However, one can say that the disturbance created was momentary, not very intense, not very frequent and, when necessary, followed by total reassurance and resolution. Nevertheless, such "ethical questions" are also an aspect of scientific epistemology and, in fact, become a part of each social scientist's field of preoccupation whenever he or she is studying human beings and not rats.

9. Teachers' Training Courses organized by Pino De Sario, director of Ecosphere (Florence, Italy) promote the use of "partial mirroring" as a means of improving contact within teacher-student relationships. 

©Jerome Liss, 1999


  1. Bandler, R. and Grinder, J., (1979) Frogs Into Princes, Moab, Utah, Real People Press. 
  2. Boadella, David, (1987) Lifestreams: An Introduction to Biosynthesis, London, Routledge and Kegan Pubs.
  3. Dore, J., (1975) "Holophrases, Speech Acts and Language Universals," in Journal of Child Psychology, No. 2, pp. 21-40.
  4. Hempel, Carl, (1952) Fundamentals of Concept Formation in Empirical Science, International Encyclopedia of Unified Sciences, Vol. II, No. 7, Chicago, University of Chicago Press. 
  5. Farrell, B.A., (1986) The Standing of Psychoanalysis, Oxford, Oxford University Press.
  6. Kurtz, Ron, (1983) Hakomi Therapy, Boulder, Boulder Press.
  7. Laing, R.D. (1967) The Politics of Experience, London, Penguin Books. 
  8. Liss, Jerome, (1986) "Epistemology of Science and Clinical Research," ("Epistemologia della Scienza e Ricerca Clinica,") Cap. XIV, di (Boadella, D. e Liss, J.) La Psicoterapia del Corpo, Roma, Astrolabio, 1986. (also in Psicologia Clinica, (editor Prof. M. Bertini), No. 2, 1983, pp.143-163.) 
  9. Liss, Jerome and Stupiggia, Maurizio (editors), (1995) La Terapia Biosistemica, Milano, Ed. Franco Angeli, (published as La Thirapie Biosystimique, Paris, Ed. Hommes et Perspectives, 1997).
  10. Massimi, F. and Inghillesi, P., (1987) L'Esperienza Quotidiana, Milano, Ed. Franco Angeli.
  11. Papousek, H. and Papousek, M., (1979) "Early ontogeny of human social interaction: Its biological roots and social dimensions," in M. von Cranach, et al, (editors) Human Ethology: Claims and Limits of a New Discipline, Cambridge, Cambridge Univ. Press.
  12. Stern, Daniel, (1985) The Interpersonal World of the Child, New York, Basic Books.
  13. Trevarthan, C. and Hubley, P., (1978) "Secondary Intersubjectivity: Confidence, Confiders and Acts of Meaning in the First Year," in Lock, A. (editor) Action, Gesture and Symbol, New York, Academic Press.