The analogicalpropositional debate

Ho'oponopono Certification

About Negative Thinking

Get Instant Access

The debate between proponents of analogical and propositional mental representations is one of the oldest in cognitive psychology. Theoreticians such as Paivio (see Paivio, 1971, 1986) have argued that both forms of representation are essential to any understanding of human cognition. In contrast, psychologists such as Pylyshyn (e.g., 1973, 1984) have proposed that all mental knowledge can be represented in propositional terms. What then do we mean by the terms analogical and propositional?

The prototypical example of an analogue representation is a picture. If we consider a colour photograph of the Taj Mahal, it is clear that structurally it resembles the real Taj Mahal. The colour of the marble is similar, as is the spatial configuration of the dome and of the minarets. In the picture we can see that the dome is in the centre of the four minarets. Such structural resemblance means that this image of the Taj Mahal is analogous to the real building. In contrast, if we consider the English sentence "The dome of the Taj Mahal is in the centre of its four minarets" we are clearly dealing with a non-analogical representation. In the sentence there is no reason why the configuration of letters "T-A-J M-A-H-A-L" should represent the famous building in India; indeed, in the original Hindi language the building is obviously represented by a completely different set of symbols. A further difference between analogical and non-analogical representations is revealed when we consider the relation between the position of the dome of the Taj Mahal and the minarets. The relation between these architectural features requires an explicit sequence of symbols in the non-analogical sentence; namely the phrase "in the centre of". In contrast, the relative positions of the minarets and dome is implicit in the picture—no explicit symbol for the relation is necessary. A further difference between the two types of representation is that the sentence can be broken down into words which in turn can be broken down into letters. Further deconstruction is not possible if one wants to retain meaning in the language, as a fraction of a letter means nothing in English. In contrast, pictures have no such discrete symbols; there is no smallest unit in the image of the Taj Mahal. Finally, the non-analogical sentence has clear rules of combination, in the form of English grammar, for the symbols it includes. Pictures, on the other hand, do not possess clear rules of combination.

Clearly we have not been talking about the format of mental representations in the preceding section but of external representations—pictures and written sentences. However, the crux of the analogical vs propositional debate among mental representation theorists is contained in this discussion of external representations. The differences between analogical and non-analogical external representations above are paralleled by the distinctions drawn by theorists such as Paivio between analogical and propositional mental representations. In contrast, other mental representation theorists such as Pylyshyn, as we have noted above, deny the validity of this distinction and propose that there are no analogue mental representations and that the external world is represented mentally by propositions. It is beyond the scope of this chapter to enter into the intricacies of the debate between these two schools of thought. An excellent blow-by-blow account is provided by Eysenck and Keane (2005) and a number of the important original source articles are reproduced in Collins and Smith (1988). The line that we will take in this chapter and those that follow is that a concept of analogical mental representations is both useful and valid when we come to try and understand emotion-related phenomena. Such representations, as we have indicated, have similar properties to the colour photograph of the Taj Mahal. That is to say, they are non-discrete, they represent things implicitly, and they have loose rules of combination. Finally, analogous representations are tied to a particular sense modality (e.g., vision in the example of the Taj Mahal). This last point is important because we endorse the view that there are several modalities of analogical mental representation that are important in understanding emotions. These are olfactory, auditory, gustatory, visual, proprioceptive, and tactile. The importance of these different forms of analogical representations is highlighted to some extent in the discussion that follows. However, the main advantages of endorsing a role for analogical representations in the emotional system are more fully explored in the chapters in the second half of the book.

So, where does this leave propositional representations? In contrast to analogical representations, propositional representations are considered to be explicit, discrete, and abstract entities. They represent beliefs, ideas, objects, and concepts, and the relations between them in a form that is not specific to any language. In other words, propositions are a kind of language of thought (e.g., Fodor, 1987), a mental lingua franca, in which different components of the mind communicate. However, an essential point about propositional representations is that although they are language-free, their entire semantic or meaning context can be captured by statements in natural language. So, for example, the propositional information that the capital of France is Paris, although not represented in the mind in English or French, is entirely captured by sentences in either language; "Paris is the capital of France", or "Paris est la capitale de France". Both sentences have the same propositional content. Given this equivalence, cognitive scientists are then faced with the problem of how to characterise these non-language-specific representations. This problem has traditionally been side-stepped by presenting propositions in the form of the logical system of predicate calculus. So, for example, the natural language sentence in English, "the writing is on the wall", can be characterised in propositional terms using predicate calculus as:


It is not necessary to enter into the intricacies of predicate calculus here. The essential point is the postulation of a form of mental representation—the propositional—that is able to represent ideational content in a non-language-specific way but for which the meaning can be totally captured by sentences in natural language.

So far we have talked about two forms of mental representation: analogical, which includes visual, olfactory, auditory, gustatory, proprioceptive, and tactile images; and propositional, which is a non-language-specific representation of beliefs, ideas, and concepts (NB: the term "proposition" is occasionally used in cognitive psychology to represent ideational content that is of a higher order than natural language and which can be only be expressed in natural language with a corresponding loss of meaning. It is important to emphasise that for our purposes, the former usage of the term proposition is employed throughout). Earlier in our summary of the important points to arise from Chapter 3, we emphasised the advantages of multiple levels of meaning in any analysis of mind. Consequently, because propositions refer to thoughts and beliefs that can be expressed in natural language without any corresponding loss of content, there remains a need to propose a structure that can represent higher-order ideational content which cannot be expressed in natural language; for example, the models of the self, the world, and others to which we referred in the previous section on the content of the mind. In cognitive psychology concepts such as schemas (e.g., Bartlett, 1932; Piaget, 1954; Rumelhart & Ortony, 1977), scripts (Schank & Abelson, 1977), frames (Minsky, 1975), memory organisation packets (MOPS) and thematic organisation packets (TOPS) (Schank, 1982), mental models (Johnson-Laird, 1983), the implicational level of interacting cognitive subsystems (ICS; Teasdale, 1999; Teasdale & Barnard, 1993; see Chapter 3), and thematic abstraction units (TAU) (Dyer, 1983) are all concerned with the representation of knowledge at higher levels than that of verbally expressible propositional concepts. Again, it is beyond the scope of this chapter to discuss the pros and cons of these varieties of higher-level meaning representations. For our purposes we shall borrow a term from Teasdale and Barnard (1993) and refer to these higher-order representations of ideational content generically as schematic models.

We discussed briefly the distinction between propositional representations and these schematic model (or implicational) representations in our discussion of ICS (Teasdale & Barnard, 1993) in Chapter 3. Teasdale and Barnard captured some of the essence of the two levels of meaning by comparing the following extract from Keats' "La Belle Dame Sans Merci":

Passage one (poetic version: implicational meaning):

O what can ail thee, knight-at-arms, Alone and palely loitering?

The sedge has wither'd from the lake, And no birds sing.

Passage two (prose version: propositional meaning):

What is the matter, armed old-fashioned soldier,

Standing by yourself and doing nothing with a pallid expression?

The reed-like plants have decomposed by the lake And there are not any birds singing.

As Teasdale and Barnard note: "The implicational meaning or 'sense' conveyed by the two versions is very different. The poetic form conveys a sense of melancholy, emptiness, and abandonment that is largely lost in the much more 'matter of fact' tone of the prose version" (p. 73). This example from Teasdale and Barnard captures the sense of evoked feelings and understanding which is carried by schematic model information. The second passage contains little more than straightforward propositional information. In contrast, the first passage unlocks whole areas of higher-order meaning.

A less poetic and slightly different illustration of the schematic model level of representation is provided by Schank and Abelson (1977) in their discussion of restaurant "scripts". Schank and Abelson argue that for commonly encountered situations such as going to a restaurant we have schematic representations, or "scripts" as they call them, of the order and type of events that might happen. These scripts help us to organise and process the information with which we come into contact in the restaurant situation (see Table 5.3).

Schank and Abelson suggest that the restaurant "script" is constructed by abstracting large amounts of propositional information and combining it to form a

Table 5.3 The components and actions of the restaurant script

Script name


Specific action

Eating at a restaurant


Walk into restaurant Search for table Decide where to sit down Go to table Sit down

Look at menu

Waiter arrives

Order drinks

Choose food

Waiter arrives

Give orders to waiter

Waiter takes order to cook

Wait, talk

Cook prepares food


Waiter delivers drinks Cook gives food to waiter Waiter delivers food Customer eats Talk


Waiter writes bill Waiter delivers bill Customer examines bill Calculate tip Leave tip Gather belongings Pay bill

Leave restaurant

Based on Schank and Abelson, 1977.

Based on Schank and Abelson, 1977.

schematic whole, which is larger than its constituent propositional parts. However, it could be argued that a more sophisticated schematic model (or script) of restaurants would include analogical information about visual appearance, smells, sounds, and also the feelings associated with going to a restaurant and so on. This would give the schematic model of the restaurant more of the properties of the poem in the example from Teasdale and Barnard (1993). On these lines, it is important to emphasise that schematic models are not mere collections of propositional beliefs. Rather they are abstracted across all of the levels of the system including the analogical. In this sense, schematic models differ from, for example, Beck's schemas (see Chapter 4) which Beck refers to as integrated collections of propositional beliefs.

The two examples we have borrowed in order to illustrate the distinction between the propositional level of meaning and the schematic model level have an important difference. In the poetry example of Teasdale and Barnard we are constructing a schematic model "on line" as we read the poem and our awareness of this schematic model is what provides us with the feelings of poignancy and sadness that the poem evokes. In contrast, Schank and Abelson's restaurant script describes a schematic representation in long-term memory. This script is activated and utilised whenever restaurant-related information is encountered. Clearly, the schematic model level of meaning usually functions in a combination of these two ways. Stored schematic representations in long-term memory are called up and combined with analogue and propositional information to build an on-line model of the current situation.

It is becoming increasingly clear that such schematic models are central to an understanding of emotions and, as with propositional representations, it is important to decide on a notational system to represent them in the discussions that follow. Any notational system in natural language will, by definition, suffer from a significant loss of schematic-level meaning. However, with this caveat in mind, we propose the following notation:


to represent the schematic models of the world as reasonably safe and as one in which bad things generally do not happen to me, rather than a propositional belief that this is the case.

To sum up, in this section we have discussed three formats of mental representation: analogical representations, including visual, olfactory, gustatory, proprioceptive, tactile, and auditory "images"; propositional representations, which represent the ideational content of the mind in a non-language-specific way but which can be directly translated into natural language; and schematic models, which contain higher-order abstracted information about the self, world, and others that cannot be fully expressed in natural language. Later in the chapter we discuss one further representational format—the associative. However, in the next section we address briefly how the content of the mind is captured in the three different representational formats that we have introduced so far, before describing the first of two proposed routes to the generation of emotions that utilises this representational scheme.

Was this article helpful?

0 0
Empowered Happiness Bible

Empowered Happiness Bible

Get All The Support And Guidance You Need To Be A Success At Being Happy. This Book Is One Of The Most Valuable Resources In The World When It Comes To Everything You Need To Know To

Get My Free Ebook

Post a comment