Intentionality, as Brentano originally introduced the term in modern philosophy, was meant to provide a distinctive characteristic definitively separating the mental from the physical.(1) Mental states have an intrinsic relationship to an object, to that which they are "about." Physical entities just are what they are, they cannot, by their very essence, refer to anything, they have no "outreach", as one might put it. Mental states have, as it were, an incomplete essence, they cannot exist at all unless they are completed by something other than themselves, their object. Brentano's position is opposed to all theories which represent the mental as only extrinsically related to the world, that is, to all theories in which mental states are themselves self-sufficient for their own existence and only secondarily relate to the world by means of something external to their nature, e.g., neurological causation, divine intervention, or pre-established harmony. In these later cases, any mental act whatsoever could be related to any object, or indeed to none, for the relation is external to the nature of the act, it is superimposed on it by outside forces. Brentano's point is that a mental act has, by its very essence, an Intentional object without which it would not be a mental act. It would therefore appear that since causality is an external relationship which could in principle relate any two things regardless of their nature, the Intentional relation between an act and its object cannot be a causal relation.
Husserl borrowed Brentano's concept of Intentionality, but used it in a "transcendental" scheme: Intentionality is a process of constitution, of meaning-giving, and so must be radically distinguished from that which has been given meaning, the object. Physical things, objects of nature, and causal processes, as constituted entities, are dependent for their meaning on a transcendental Intentionality which cannot therefore, on pain of circularity, be reduced to, or explained by, any physical process. Transcendentalism, then, involves a dualism (though not a dualism of substance) between Intentionality on the one hand and nature and causality on the other. It is this dichotomy that any attempt at a naturalistic theory of Intentionality must overcome.(2)
John Searle attempts, in Intentionality, to develop such a naturalistic
theory.(3) In what follows I will present
briefly the central points of his theory of Intentionality. In my discussion
of the theory, I will look at three issues: The nature of the mental state
to which Searle attributes Intentionality; his notion of Intentional causation
by which he hopes to bridge the transcendentalist gap between the Intentional
and the causal; and the relationship of the mental structures of Intentionality
to the physical structures of the brain, a relationship that Searle understands
OUTLINE OF SEARLE'S POSITION
John Searle accepts Brentano's basic definition of Intentionality as the logical property of being about an object, though he usually substitutes states of affairs for objects. Borrowing from Frege's notion of Sinn, Searle maintains that every Intentional state has an Intentional content which determines conditions of satisfaction for the state. It is through the Intentional content that the Intentional state is linked to its object. The state can then be said to "represent" the state of affairs which satisfy these conditions, though this is a purely logical characteristic implying no images or "representations": Intentional states "represent" only in the sense that language can be said to "represent." Each state also has a psychological mode which determines the direction of fit: mind to world or world to mind. In belief, for example, validity (in this case truth) is achieved when the mind matches the world; in a valid (i.e., successful) desire the world must come to match the mind. The conditions of satisfaction for many Intentional states include a self-referential clause; perception, for example, has as part of its very meaning that it be a state caused by the object represented in it. This aspect of the Intentional content remains even in hallucinatory experiences in which the conditions of satisfaction are not met.
For example, Searle would analyze my Intentional state of seeing a car as:
In Searle's theory it will be useful for us to distinguish four relationships between Intentional states and reality: Language to object, Intentional state to object, Intentional state to psychological infrastructure, and Intentional state to neurological infrastructure. I will explain the first three in turn, and then offer some criticisms, leaving consideration of the fourth relationship, based on a special kind of causality, to a later section of the paper.
Language relates to reality, in Searle's approach, by speakers so relating it in their speech acts. Speakers use sentences to represent the meaning they wish to express (22, 197). The understanding of linguistic meaning therefore depends on our analysis of mental Intentional states, and so the relationship of language to reality reduces to a special case of the relationship of mind to the world.
Mental Intentional states, according to Searle, do not relate to reality in the same way that words do. We cannot use a belief, for example, in one way rather than another, for its Intentional content determines its own conditions of satisfaction. To say otherwise would require us to invent an infinite regress of mysterious agents, homunculi, each using the representational states of the lower homunculi to mean something (21). The buck must stop somewhere. Searle stops it at the first Intentional state, maintaining that its Intentional content logically and intrinsically determines its own conditions of satisfaction. A belief that the moon is red cannot be used to mean the cat is black, even though the sentence "the moon is red" could, if we chose to use it that way.
On the other hand, Intentional states, Searle maintains, are only empirically linked to the psychological experiences which embody them. Perception, for example, involves "perceptual experiences", though it is to be distinguished from them (38). Perception, as an Intentional state, has propositional content, i.e., Intentional content, but it also involves conscious mental acts. These must not be confused: "One can say of a visual experience that it has a certain temporal duration or that it is pleasant or unpleasant, but these properties of the experience are not to be confused with its Intentional content" (43). "The material object can only be the object of visual perception because the perception has an Intentional content, and the vehicle of the Intentional content is a visual experience" (61, cf. 58). Searle considers the arguments of those who deny the existence of such psychological experiences, and even presents pathological examples of "blind sight" in which patients claiming to have no visual experience of an object can nonetheless "see" it in some way, for they can answer correctly questions about it (47). Yet he remains convinced that perceptual Intentional states, at least, are in fact related to "experiences", even though the relationship is not logically required.
Searle summaries his position as follows:
I want first, on the basis of this citation, to look at Searle's notion of an Intentional state before examining the issue of causality. The above citation seems very clear, but unfortunately this clarity is not representative of all of Searle's book, which, on these points, suffers from numerous slippages of sense. His position that "an object is referred to in virtue of satisfying an Intentional content" (222), implies that it is not in virtue of its mode of realization that an Intentional state refers to an object. Searle, however, not only fails to say this explicitly, but, despite his distinction between perception and visual experience, continually speaks of visual (and other) experience itself as presenting a state of affairs (46), as being satisfied or unsatisfied (50), as having an Intentional content (56), as determining conditions of satisfaction (61), and so on. This is an easy mode of expression, and often causes no difficulty, but strictly speaking it is incorrect: It is not the visual experience as a mental event which has conditions of satisfaction, but the perceptual state that possesses these logical properties, and it does so, not in virtue of the perception involving a visual experience, but in virtue of it having an Intentional content. Having attributed the logical properties of Intentionality to "perception," Searle goes on to speak of "perceptual experience" (45), then more loosely of "experience" without qualifying it as visual or perceptual (68). It is not surprising after these verbal shifts that he slips easily into speaking of "visual experiences" as themselves having conditions of satisfaction (76-77).
The argument that Intentional content is not "used" to refer to an object, but by its logical structure intrinsically determines its object itself, seems plausible. It cannot, however, be extended in this surreptitious manner to include a claim that a visual experience (as opposed to a perceptual state) is logically related to its object. Visual experiences are only empirically related to Intentional content. We could imagine a non-human creature whose perception of a tree, while having the same Intentional content, is realized in a radically different psychological structure, or perhaps in none at all as the blind sight case suggests. As psychological events, visual experiences are neutral with respect to reference, as neutral as words, and only when they realize specific Intentional contents do they take on logical properties as satisfied or unsatisfied, fulfilled or unfulfilled, true or false.
It is therefore misleading to say that "there is no way to describe my visual experience without saying what it is an experience of" (43). A visual experience is not itself, strictly, of anything, it just is. Only when it takes on logical properties, only when it acquires an Intentional content can it be about an object. Perception, and other Intentional states logically determine conditions of satisfaction and thereby refer to an object. Visual experiences do so only secondarily, empirically, contingently, in so far as they are related to Intentional content by what Searle calls "realization."
But what is the relationship of realization? That is, how can a psychological process come to have an Intentional content? How can such a state take on the logical property of having conditions of satisfaction? The link from content to psychological state needs explanation: Why this state with that content? The answer that works for words, "because we give it that content, we use it that way," fails here for two reasons. First, we would be at a loss to know what to do if asked to use one of our psychological states to intend something different; indeed we have no awareness of using psychological states for anything. Secondly, this approach would lead us back into the infinite homunculi difficulty, as Searle himself explains clearly (21). So realization cannot be explicated by the notion of "use."
I submit that Searle cannot tell us what realization is from within his naturalistic framework, for he lacks the notion of transcendental subject. The closest he comes to the concept is his notion of "homunculus," which appears to be but a set of psychological states. But psychological states are objects like any other. They can, for example, be the objects of Intentional states. They would therefore seem, like words, to be available for use in expressing whatever one might wish. But by whom? Not everything can be "used by", for we must finally reach "something that does the using and so is not used," i.e., the subject. Searle is caught in a bind. On the one hand he accepts that objects which can be used to express meaning, such as words, depend for their meaning on the user, the "meaner" of the meanings, and that no infinite regress can be permitted along this line. On the other hand, the only possible candidate in his system for stopping point is a psychological state or set of states, but such states seem themselves but objects to be used.
A way out might be found in Searle insistance that his concept of Intentional content is purely logical, that it has no "ontological" status. If he could associate a purely logical subject, something like a transcendental ego, with this content, his search for a ultimate stopping point would be accomplished. An Intentional content, as he sees, has no being other than its meaning; there is no contingency between its being and its meaning, and so there is no danger of regress. Realization, then, would be the use by such a transcendental subject of psychological states to express a meaning, a relation analogous to the use of words.
Searle's naturalism, however, prohibits this solution. Only natural, psychological states can function as users, and so the realization of Intentional content in psychological states is left incomprehensible. Or rather, the gap between them is covered over by a verbal equivocation.
To summarize, then, Searle introduces two different relations. There is the relation between Intentional content and reality in the sense of the object intended, and this link of mind to reality is, according to Searle, logical. There is also the link of the content to reality in the sense of the psychological state in which it is realized, and this, Searle claims, is contingent. By a slippage of meaning, Searle thinks that he has been able to show a logical and Intentional relation between psychological state and intended object, but his naturalism makes the crucial link, the concept of realization, incomprehensible.
One form of the "transcendental gap" in Husserl is the incomparable
nature of transcendental Intentionality and any mode of objective being,
including psychological states. The very first step in any naturalistic
theory is the capturing of the logical in the psychological, a preliminary
to the attempt to relate both to the physical. I will examine the second
step later; what I am suggesting here is that Searle has not managed to
achieve even this first step. The relation between Intentional state and
psychological state, or, as Husserl might put it, between the transcendental
and the empirical, is as mysterious as ever.
One of Searle's most interesting points is his theory of the Intentional relationship between Intentional state and its object in the case of perception and action. In both these cases, as we have seen, he maintains that a self-referential causal element is involved. It is part of the Intentional content of perception that the experience be caused by the object perceived. Similarly, the Intentional content of an action includes a requirement that the desired result be caused by the action itself. This is the notion of "Intentional causation." Self-referential causality permits him to link the intrinsically general sense of the Intentional content to the singular, concrete nature of the actual world grasped or modified in a particular act. The causal relation is then subsumed as part of the very content or meaning of the Intentional state, while not ceasing to be causal. This theory of "Intentional causation" would be a brilliant synthesis if it were successful. In particular, as he points out himself, it would be a major step towards "naturalizing Intentionality" and would enable him to reject the view that Intentionality is transcendental, beyond the natural world (112). Surprisingly, however, the direct target of his arguments is not a transcendentalist such as Husserl, but Hume.
What does Searle mean by Intentional causation? He maintains that Intentional causation is but a sub-species of causation in general. Positively the notion of cause is difficult to define; there is little we can say about it. Since we all perceive and act, however, we all have the direct experience of "making something happen"(123). If I did not experience the car as making my perceptual state happen, then my experience would not be one of "perceiving the car." The general notion of cause is derived from this primitive experience. Searle contrasts the notion of cause that we gain in this way with the traditional, i.e., Humean, account on three central points: causation need not be unobservable; causation is not based on universal regularity; and the causal relation is not to be opposed to logical relations. I will look at the first two only briefly and the third in more detail.
Hume claims we can observe only a chain of events, never a causal link between them. Against this, Searle maintains that in action and perception we experience the causal relation itself; this is Intentional causation on which our general notion of cause is based. This "experience", however, is of a peculiar sort: A cause is not "observed" or "seen" as the object of an experience (49, 124), it is given as an essential part of the Intentional content of the experience. While Searle explicitly rejects the Kantian theory that causality is an a priori concept, his own position is closer to it than he seems ready to admit. He claims Intentional causation is not discovered in an empirical, contingent manner like a table, or gravitation; it is intrinsic to the innate or logical structure of perception and action. As Kant might put it, cause is experiential in that it is an a priori form of our experience, not in the sense that it is a contingent element of experience. Although they may differ on many other points, Kant and Searle agree that we do not discover causality as an empirical feature of the natural or psychological worlds. It is a logical feature inherent in perception and action.
Secondly, despite Hume's analysis, regularity is not implied by the notion of Intentional causation, according to Searle. My thirst causes me to drink this time, but it might not next time; it's up to me (118). I think this is a weak argument on two grounds. First, it fails to mention, much less discuss, the traditional distinction between reason and cause: If thirst is a reason for my action and not a cause of it, then Searle's position would be untenable. Second, even if thirst is a cause, it may not be a sufficient cause, in the sense that it could not alone explain the occurrence of the effect. If the same condition of thirst occured a second time without subsequent drinking, the difference between the two cases would remain unexplained, and one would be led to presume that some other, as yet undetected factor was at work. A claim that one had found the sufficient cause which explains why the event occured would involve an assumption that the effect would invariably follow the cause in all future cases. I think that the most that Searle could claim here is that there are loose, informal uses of the term "cause" in which people haven't thought through the implications of what they are saying. As the term is used scientifically, regularity is implied (though it may not be the essence, as Hume thought).
Searle's arguments against regularity and unobservability, however much they may undermine the Humean account of causality, do not bear significantly on the transcendentalist's dichotomy of Intentionality and causality. The rejection of the non-observability of causation, even if correct, does not bridge the gap. One who believes that causation is always natural will not be forced to change their mind if they agree that "making things happen" rather than regularity is the essence of causality. The basic issue facing an naturalist is that meanings, ideas, Intentional contents, and so on, appear to take part in relationships of a radically different order than those between objects of the natural world. The difficulty crops up in one form even in Hume, though he is hardly a transcendentalist, and appears in his contrast of the external, contingent nature of causal relations with the necessary nature of internal, logical relations. It is in this form that Searle attacks the issue.
Searle argues that, despite Hume, an object which is the (Intentional) cause of a perception is logically or internally, rather than contingently, related to the Intentional content of that perception. The cause of my perception of the table must, logically, be the very table which fulfills the conditions of satisfaction laid down by the Intentional content of the perception. Hence the causal relationship, at least in the case of Intentional causation, is not a purely empirical relationship, but involves an element of meaning. Intentional causation gets beyond Hume's logical/empirical dichotomy and so bridges what I have called the transcendental gap. This is a very interesting argument, but I think it is unsound for the following reasons.
Searle himself points out a "harmless" ambiguity in the notion of "conditions of satisfaction." It may refer to a "requirement", i.e., that which would have to be met if my perception or belief were to be true. The requirement is internal to the Intentional content and remains the same whether the perception is hallucinatory or veridical. But "conditions of satisfaction" may also refer to the state of affairs in the world which actually satisfies the Intentional state, if the state is true, the "thing required" (13). In a similar way, I think we must distinguish between the "cause-as-requirement," which is determined by the Intentional content of, say, a perception, and the "actual-cause" of that experience. In a veridical perception of a table these coincide, for part of the Intentional content is that the experience be caused (cause-as-requirement) by the table, and it in fact is caused (actual-cause) by it. In an hallucination of a table the cause-as-requirement is still the table, but the actual-cause is something else, for example, a drug or electric brain-probe.
Searle argues that we can get the direct "experience" (in his special sense) of causation, and hence the general notion of causality, from what I have called the cause-as-requirement even if it does not coincide with the actual-cause, and I am willing to grant him this point (130). The argument does not however help him to establish that there is a logical relation between the Intentional content and the cause of a particular perception.
There is indeed a logical relation between the fact that a perceptual experience is "of a table" and the Intentional content having "the table" as its cause-as-requirement. But since there is no guarantee that the conditions of satisfaction of the Intentional content are being met (it may be an hallucination), there is no logical necessity that the actual-cause be the table. Let us grant Searle's argument that it is a logical necessity that in a veridical perception the actual-cause must coincide with the cause-as-requirement. Since veridicality is itself a purely empirical relation, this does not establish his claim that there is a logical link between Intentional content and actual-cause. The perceptual experience "of a table" might be caused by an elephant. It would not thereby cease to be the perceptual experience "of a table," which it must logically be. It is simply false, hallucinatory. In veridical experiences the cause-as-requirement and the actual-cause coincide, but that the experience is a veridical one is, on Searle's theory, purely contingent. The truth-value of an experience may depend logically on the conditions of satisfaction, among other things, but this does not imply that the actual-cause is logically dependent on them. Hence Searle has not shown that there is an internal relation between the Intentional content and the actual-cause.
So nothing in the Humean dichotomy between logical and causal is undermined by Searle's analysis. The actual-causal relation may still be completely external, contingent, unpredictable by reason. The relation of Intentional content to cause-as-requirement may still be purely rational, logical, internal. The value of the perception, i.e., its veracity, is logically dependent on the contingent matching of cause-as-requirement and actual-cause, but why would a Humean object to that? Yet if Searle cannot make his case that Intentional causation is a logically necessary relationship, then his main argument, not only against Hume, but against the transcendentalist position, collapses.
Hence I submit that Searle's theory of Intentional causation has failed to bridge the transcendentalist's gap between Intentionality and natural causality, which he claims was the main object of his strategy. His arguments against regularity and unobservability interpretations of causality, even if correct, are of little relevance to the issue. The relevant argument, that against the purely external nature of Intentional causal relations, founders on an ambiguity he indicated himself but thought harmless.
Indeed Searle himself is forced to accept the dichotomy in his later
chapter on proper names. He is opposed, for reasons that need not concern
us here, to the causalist theory of proper names. His own theory involves
ostension, and so perception, and it therefore also depends on causation.
"But it is Intentional causation, internal to the perceptual content, it
is useless to the causal theorist in his effort to give an external causal
account of the relation of name to object. ... To give the ostensive definition
the perceiver had to perceive the object and that involves more than the
physical impact of the object on the nervous system" (235). Here Searle
is himself acknowledging that the relationship between the Intentional
content and its object is different than the relationship between objects
in the world, in effect reinstating the transcendental dichotomy.
INTERLEVEL CAUSATION AND THE MIND-BODY PROBLEM
Finally, let us turn to the relationship between Intentional states and their physical infrastructure, that is, to the mind-body problem. The naturalist position must not only attribute logical and Intentional properties to psychological states, but must relate these psychological states to neurological processes. Searle, in the Epilogue to his book, offers an account of the relation of mind and body compatible with his naturalistic theory of Intentionality. He contrasts his theory with two traditional ones, interactionism and the identity theory. The interactionists maintain that mental states are caused by neural events. Identity theorists claim that mental states and physical states are the same thing under different aspects or descriptions, and so one cannot cause the other; the most one can say is that the mental is realized in the physical. Searle wants it both ways: Mental states are realized in and also cause, and are caused by, physical states. To achieve this he must, once again, modify the received notion of causality.
He conceives of the mind and mental states as higher level biological states of the organism, similar in their logical status to the functionings of the kidney. His thesis is that "two phenomena can be related by both causation and realization provided that they are so at different levels of description" (266). Thus the liquidity of water is both caused by and realized in its molecular structure. In an engine, the rise in temperature of the gases is caused by and realized in the movement of electrons across the spark-plug gap, and the resultant explosion in the cylinder is caused by and realized in the oxidation of the molecules. Thirst, as a mental state, is caused by and realized in certain neural events, probably in the hypothalamus. (We should note that Searle is speaking here of external, physical causality, and makes no reference to his doctrine of Intentional causation in this context, except insofar as all our notions of causality are derived from our "experience" of Intentional causation.)
I have no problem with liquidity being realized in the molecular structure, but I think that Searle is wrong in claiming that it is caused by this structure. Searle claims that "if...we alter the molecular structure we cause the surface features to change" (266). This is nonsense. It is like the claim that we can cause someone to be married by making him a spouse; these are two descriptions of the same event, so one cannot be the cause of the other. We can heat the ice, or pass an electric current through it, and it will melt, but this is causality on the surface level only. We can say that a stream of electrons changes the molecular movements, and this is causality on the microlevel. I can even accept, as does Searle, to say that we can heat the ice and so cause the molecular structure to change. But the change of molecular structure cannot cause the melting, it is the melting.
This should be particularly evident if we realize that the above citation could just as well have been: "if we alter the surface features we cause the molecular structure to change." Could a causal relationship be symmetrical in this way? Even if Searle were right that the causal relation is not essentially unobservable or regular, some defining characteristics must remain. Should we say that, as a minimum requirement, the cause must precede the effect?
Searle might object to this by calling on a distinction he makes between "causings" and other causal relations. "The billiard ball is gravitationally attracted to the center of the earth" is not, he claims, a relation between events, and so is not a causing, but it is nonetheless a causal relation (116). Perhaps only causings require an antecedence relation between a cause and an effect. But this objection would lead nowhere, for, I maintain, causal relations which are not causings are not explanatory either. Only events, not objects, can be explained. The presence of an object, or its possession of a property, are events explainable by causings, but an object's causal relations (if they are not causings) do not explain anything. Clearly, however, molecular structure is meant to explain liquidity. So Searle could not escape in this way the requirement that a cause must precede its effect.
But could he not argue that even causings do not require antecedence? Could an object's falling not be explained as caused by the action of gravity which occurs simultaneously with the fall? Maybe; but we still could not say that the falling caused the gravity. Some priority must be given to the cause, even if it is not a temporal priority. The cause must be independent of the effect; the cause must be independently variable, or, at the very least, there must be an asymmetry which allows us to distinguish cause and effect. If the relationship between two events is symmetrical, then it is not a causal relation.
Hence, since melting the ice involves changing the molecular structure just as certainly as changing the structure involves melting the ice, the relationship is symmetrical and therefore not a causal one.
We might be misled by the pragmatic fact that, in the case of human intervention, we more usually describe what we do as melting the ice, and so might be led to think of the changing molecular structure as less direct. But referring to the change in molecular structure is at most an inappropriate mode of description in the context of action; it is not a reference to a separate event brought about indirectly by means of melting the ice. We could also be misled by the fact that the molecules explain the liquidity and not vice versa, but this is a different matter. Many explanations are non-causal. The motion of the earth can be explained by the premises that all planets move in ellipses, and the earth is a planet, but neither premise offers a cause for that motion. To explain the surface properties of water by reference to the microlevel, or to explain thirst neurophysiologically, does not necessarily involve a claim that the lower level causes the higher.
It therefore seems that Searle's claim that Intentional states can be
causally related to the brain states in which they are realized must be
rejected. So even if Searle had been able to adequately explain the relationship
between Intentional, mental and/or psychological states, the final naturalistic
step of linking them to their physical realization is still incomplete.
Naturalizing Intentionality is not, then, as easy a task as Searle might have hoped. Some of the difficulty lies in the lack of a clear definition of the project. On the one hand he wants to treat Intentional states as biological functions, like those of the stomach or kidney (15), which exist in the physical world and can be investigated by the methods of the physical ("serious") sciences (263). On the other hand, he insists, "there really are such things as intrinsic mental phenomena which cannot be reduced to something else or eliminated by some kind of redefinition" (262). So his project is not one of reduction. He is clear only that naturalism involves the rejection of transcendentalism.
At the very minimum, however, we could expect his project to explicate the relationships between Intentional states and reality. Three relationships are crucial to this endeavour. The logical structure of Intentionality must be realized in psychological states; it must be related to the object Intended; and it must be related to brain states. Searle seems oblivious to the first relationship, escaping the difficulties by a shifting use of terms. The second is explained by the use of "Intentional causation", his major inspiration, but this notion, I have tried to show, fails to overcome the transcendental dichotomy. The third he explains as an ordinary causal relationship, but only, I have suggested, by abusing the notion of cause.
Searle's logical analysis of Intentionality is interesting and promising,
but the supporting ontological framework of mental, psychological and brain
states, and the causal relationships between them are still problematic.
In particular, the transcendentalist's radical distinction between Intentionality
and objects of any kind has not been overcome. Intentionality may be naturalizable,
but has not, as yet, been naturalized.
1. This paper grew out of discussions at a seminar of the McGill Philosophy Department. I gratefully acknowledge the hospitality and inspiration of the participants.
2. This is an oversimplification. For Husserl, the psychical is an much a constituted, objective realm as the physical; Intentionality is attributed to the Transcendental Ego, not the empirical ego. Searle ignores this point and attributes Intentionality to loosely defined "mental states." He is therefore using Intentionality more in Brentano's sense than Husserl's. I cannot, however, discuss this problem here.
3. John Searle, Intentionality (Cambridge: Cambridge University Press, 1983). All in-text references are to this work.
4. Modelled on Chapter 2, especially p. 48.