Interactivist Summer Institute

July 22 - 26, 2003

Copenhagen


Metaphor in Architectural Design Process from an Interactivist Framework



MOHAMAD RADZI MUSTAFA
LANGUAGE DEPARTMENT
MARA UNIVERSITY OF TECHNOLOGY
PERAK BRANCH
MALAYSIA


INTERACTIVIST SUMMER INSTITUTE CONFERENCE

SUBMISSION OF PAPER

MAJOR THEME: METAPHOR
PAPER TITLE: Metaphor in Architectural Design Process from an Interactivist Framework
AUTHOR: MOHAMAD RADZI MUSTAFA
ADDRESS: MARA University of Technology Perak Branch
32600 Bota
Perak
Malaysia

PHONE: 05-3742208 (O)
FAX: 05-3712171
E-MAIL: radzi013@perak.uitm.edu.my


ABSTRACT
This paper suggests a framework for metaphor from an Interactivist position. I shall argue that metaphor is a specialized system activity of the brain that is manifested at certain levels of representationality described by Bickhard (1999). The term ‘role’ is used to describe how it displays itself at each level of representation. Hence, there are three roles of metaphor that underlie human cognitive processes: as a vehicle for categorisation; for comparison in illuminating understanding; and for creation of new knowledge. In relation to architecture students undergoing a basic design course, I shall suggest that their thought processes tend to display metaphor’s first role - the act of categorisation. Established designers instead display the last role. It is therefore imperative that teachers of architecture education should train students to display the thought processes implicit in the second and third roles of metaphor.


NOTE ON AUTHOR
MOHAMAD RADZI MUSTAFA obtained B. Arts (Hons) Architecture from Sheffield University, UK and MSc TESOL from Stirling University, UK. He is currently with the Language Department, MARA University of Technology, Perak Branch, Malaysia. He teaches Linguistics, Architectural Design and Proficiency English at the university.

Issues

In teaching the basic architectural design course, a common challenge that I encounter regards students interpretation of felt experience using models to convey this felt experience. As part of the basic design programme, students are given an object to study, for example, a biological organism like fish or beetle, or artifacts like bird nests or weaponry. A recurring problem that I encounter is students’ type of interpretation inherent in the model making their felt experience of these objects of study. On most occasions in their design projects, when students are asked to construct a model of their experience of a particular phenomenon (e.g. their felt experience along a particular stretch of townscape) or understanding of a chosen organism (such as fish or beetle), in modelling this felt experience or understanding, there is always a tension between what I termed symbolic representations and ‘functional interactive properties’ which I shall shorten it to functional representations.

To elaborate, when students experience a part of townscape that is by the seaside, they would place water in their model to symbolize the sea instead of say, building their model to represent change in form from soft and flowing (sea) to hard and static (land) to represent interactive properties between the students and the environment. They are more aware of the objects of water and land instead of the change of landscape. If they were to experience a place full of fish, their model would be symbolised by materials in the form of fish scales and sometimes even in the shape of a fish. They could, instead, represent this experience with the movements the fish execute which would be the consequence arising out of interaction between them, the fish and its environment.

I like to sugest that this tension between the use of representations such as water, fish scales, which symbolise the phenomena experienced (sea, fish), I call these symbolic representations, instead of using representations based on ‘functional interactive properties’ that denote (in non-symbolic manner not visually dependent) soft against hard, flowing against static, or movement is because students are more aware of objects in their environment especially if these objects are already part of their vocabulary instead of the interactive properties of these objects. The recognition of objects based on vocabulary seems to hijack students experience of his/her surrounding. To say there is a tension between symbolic and interactive interpretation is not really right because in the main, the symbolic interpretations would be the norm. However, in the teaching of design, the non-symbolic interactive experience has as much, if not more, value because they involve a variety of senses such as touch, smell, taste, and sound. Whereas symbolic representations tend to rely on the sense of sight. This sense of sight alone may give rise to the other four senses in the imagination, but to be able to experience directly would be even better. It is my contention that the non-symbolic interactive interpretation should be sought from the students.

Therefore, the central thesis of this paper will address the question why symbolic representations occur prominently in architecture students interpretation in model making compared to representations based on functional representations? The assumption will be that in experiencing their environment, students tend to be more aware of objects instead of potential interactive properties of this object. This tendency will then be transferred to students model making. I shall propose an answer to this question from the framework of Bickhard’s (1998b) interactivism. This paper shall be organized along the following outline. In the first section, I shall differetiate two key concepts: symbolic representations and those based on functional interactive properties which I shall now on refer to them as functional representations. In the second section, I shall propose how the concept(s) of metaphor from an interactivist framework can elucidate the tension between architecture students’ symbolic representations and functional representations. In this section I will offer three roles of metaphor and suggest that each role correspond to a particular way of interacting with objects in our environment. Lastly, I shall conclude that symbolic representations are more prominent among students who do basic architectural design course which reflect their status as novice designers.


1.0 Symbolic and Functional Representations

Two key terms are important in the understanding of the issues involved. These are: 1) symbolic representations; 2) functional representations.

Symbolic representations

This type of representations would be understood from the interactivist (Bickhard 1998b) framework. In this research, symbolic representation is defined as that indicated by the level of knowing which Bickhard (ibid.) pointed out can be represented in the form of canonical encodings (p.18). Following this, symbolic representation can be looked at as being indicated by the level of knowing which Bickhard (ibid.) calls the level of ‘learning and learning to learn’. It is in this level that the construction of ‘stand-ins’ is required for off-line learning to take place. Like a Morse code, a stand-in “…” can represent “S” symbolically. “…” is not an exact copy of ‘S” but a representation – stand-in, for it. It is this idea of standing-in for something that a representation is a symbol. A stand–in is not original representation but subsidiary to it (Bickhard 1999: p.17).

As mentioned above, ‘water’ or ‘fish scales’ which are used by architecture students to model their experience of the sea and fish, would be looked upon as symbolic representations. ‘Water’ and ‘fish scales’ are not exact copies of sea and fish but are stand-ins for these. They are also subsidiary to the original objects since they are derived from them. I suggested earlier that the tendency to use symbolic representations in students’ design is caused by language since it is the ‘vehicle used to transport design ideas’. Language allows the parceling out the environment in such an objective manner.

How does language look like within the interactivist framework? The standard ‘transmission model’ assumes language to be a code already with representational content to be transmitted by the speaker into the mind of a listener where he/she decodes this content and hence, comes to understand what is uttered. Since the utterance is assumed to start off from the speaker already with content, this form of representation is an ‘encodingist’ one (Bickhard 1998a, 2001a). As an analogy with computing, an utterance is like a string of input. Therefore understanding in this sense is just the processing of the input. One of the shortcomings of this view is that it is not sensitive to context since the focus is on the content of the utterance (dictionary meaning) to be decoded by the listener.

The interactivist framework constitutes language and accomplishes the change and creation of representations in the listener differently from the standard ‘transmission model’. Although this is achieved differently, the end result is the same. Bickhard (1999: p20) proposes that:

…language interactions are distinguished by the objects of the interactions, by what they interact with. Specifically, I propose that utterances interact with social realities---they operate on and transform social realities of conversational flow, of discourse flow, of conventions, of institutions and activities within the framework of institutions, of relationships, and so on. Social realities are understood to be constituted as certain forms of commonality of understanding of the situation by the participants in that situation. Utterances, then, operate on and transform –just like most other interactions – but what they operate on and transform are social realities constituted as commonalities of understanding: situation conventions is the term I use for such commonalities.

In this interactivist paradigm, language ‘operates on and transforms’ situation conventions. Harnad (1982) suggested that words are categorical in the sense that a name is used to recall a discrete minimal set of invariants that is a reduced form of continuous experience. In his words, Harnad (ibid.: p.5) wrote:

…In fact, the role of the identifying response, the arbitrary, unique name, is to provide a reliable, exclusive means of access to the engram. I call such encoding “categorical” because it has well-defined category boundaries. In particular, any input continua involved have been “quantized,” in the sense that certain entire regions of continuous variation have been reduced to an effectively equivalent discrete unit. Instances have been reduced to their invariants.

Since language operates and transforms situations conventions which are further represented by language, what gets propagated will be this categorical aspect of language that Harnad (1982) pointed out. This aspect is important because it makes us think and act categorically. I argue that this categorical aspect of language allows us to see phenomena in an objective manner, i.e., focusing on the objects instead of their properties. This is what is felt to happen to my students thought processes in focusing on the objects sea and fish when transferring their thoughts into model making.

Functional representations

With regard to this paper, functional representations are to be understood as a type of representations that arises out of a system’s interaction with the objects in the environment. This interaction will have consequences for the system and will be represented implicitly as system organization. The properties of the objects in their environment that a particular system such as a human brain, recognize are then also understood to be implicitly represented. Although for humans, these properties which are implicitly represented hence, understood can be expressed verbally. Bickhard (1998b: p.6) writes that the form of implicit representation that gets constructed when an organism interacts with its environment will be “…indications of interaction potentialities…and anticipated or anticipatable interaction outcomes’. In my reading, these interaction potentialities and anticipated outcomes get their ontology from the functionality of the system as it manoeuvre itself in the environment it comes into contact with. To take a concrete example, the functional representation of a human interacting with a chair will be all the interaction potentialities such as sitting, standing, pushing, lifting, painting, and even burning the chair. But interaction is not only limited to kinesthesia. In interacting visually, the potentialities will be whatever formal attributes (texture, shape, colour, material) the chair allows us to manipulate it for example, in seeing that the chair is painted a child can scrap out (if the child has learnt this) the coat of paint with a hard object. The anticipated outcome for a steel chair visually, would be that it would be cold to the touch or hot if it lay in the hot sun for sometime.

In this interactivist framework, functional and symbolic representations are both derived from interaction. However, the latter is explicit representation operating in the realm of social conventions. Whereas functional representation is implicit constructed from the system’s interaction with the physical environment. This is a very important distinction that the author would like to draw especially for the purpose of this paper.

The functional representations in relations to sea and fish which my architecture students could have experienced in their interactions with these objects would be soft against hard, movement, fluidity, change in surface texture, etc. However, as mentioned earlier, these are the properties which escaped ‘notice’. In architecture, these functional representations are suggested to be more important than symbolic representations.

At this point, I would like to propose that the tension between symbolic representations and functional representations is due to specialized system activities of the brain which are metaphorical in nature. I will now discuss how these system activities can be seen as metaphorical and will argue that there seem to be three roles of metaphor as system activities.

2.0 Metaphor from the interactivist framework

In studying the processes of thinking, the concept of metaphor has been central (Lakoff and Johnson 1980, Lakoff 1987, Gentner and Markman 1997). Research on metaphor from a non-interactivist framework has been abundant (Black 1962, Lakoff & Johnson 1980, MacCormac 1985, Ortony 1979, Richards 1936). However, Indurkhya (1992, 1994, 1999) has written rather extensively on the subject from the perspective of interactionism.2

Bickhard (1998a) explained that interactivism is a non-encoding model of emergent representation. He proposed that mental representations emerge from the interaction between the cognitive agent and the environment. An important difference between this model and the encodingist ones is that representations in the agent are implicit. Implicit representations will allow the agent itself to detect errors in its system organization without requiring an external observer. An analogy with the computer will be that the system itself can detect errors in its programming and correct them instead of the external computer expert. This model of mental representation provides a particular view of metaphor. It should be understood as a ‘specialized system activity’, in the sense Bickhard (ibid.) uses to describe various manifestations that are familiar to us like perception for example. I would like to extend this phrase to metaphor as well.

Bickhard (1999) has also proposed that representationality should not be looked at as monolithic but comprises many levels. I would like to propose that as a specialised system activity, metaphor corresponds to three levels of representationality. The term ‘role’ is used to describe how the metaphor displays itself at each level of representation. Hence, there are three roles of metaphor that underlie human cognitive processes. It is proposed that these roles are: as a vehicle for categorisation; for comparing in illuminating understanding; and for creating new knowledge.

2.1 First Role of Metaphor: Categorisation as a specialized basic level metaphorical system activity

Categorisation sorts things. In the preface to his seminal work Women, Fire and Dangerous Things: What Categories Reveal about the Mind, Lakoff (1990: p.5) said that “There is nothing more basic than categorization to our thought, perception, action, and speech.” We are doing an act of categorising everytime we identify a particular thing as some kind of thing. This is also probably how non-human creatures go about identifying objects in their environment too. Lakoff (ibid.: p.6) also drew attention to the traditional conception of categorization which is termed ‘classical’. To be in a particular category, from the classical view, is to possess features that are necessary and sufficient for deciding what belongs in which category. To categorise things, a comparison of similarities based on the above conditions has to be made. However, Lakoff (ibid.) thinks that the classical view of category is not complex enough to explain how we categorise. If we understand classical categories as just involving the comparison of common set of features than Gentner and Markman’s (1997) proposal to reconcile similarity-based mapping with theory-based accounts is an acknowledgement of the inadequacy of the classical view of categorization in understanding the cognitive processing of the mind.

As mentioned earlier, this paper takes an interactivist view of metaphor as a special manifestation of system activity. The shortcoming of viewing metaphor from other points of view has been reviewed elsewhere (see Indurkhya 1997) so I shall not go into that again. Instead this paper proposes a way to construe various roles of metaphor as specialized activities of the system at different levels of representation. It is felt that a consequence of viewing metaphor this way is to consider the process of categorization as a metaphorical activity. How is this so?

Bickhard (1999) described his level three of representationality (interactive implicit definition and differentiation) as

…a kind of representationality that is implicit in system functioning, system environmental interaction that has already occurred (p.6)…some environments will yield the same such final state, while other environments will (or would) yield quite a different state. The final possible states of such a subsystem, then, serve as a differentiation of its class of environments (p.7)…in level three we have implicit definitions of environmental categories…(p.10)

At this level, what gets represented in the system as discussed by Bickhard is differentiation of the environments that are interacted with by the system. The outcome of such differentiations would be to group together environments that will yield the same final state after the interaction. In other words, environmental categories are differentiated by these final states. We can see that at this level of representationality that Bickhard discussed, the system activity of categorization is introduced. In interacting with our environments, our brain is already implicitly differentiating these environments and categorising them according to similarity in outcomes expressed by the final states. Representational content with regard to categorising is already implicit without us providing this content outside the system.

Harnad (1982) claimed that before we can identify a particular thing (hence put it in a particular category), our experience of instances of the thing has to be reduced to a minimal set of invariants. This will form the memory of our experience of the thing and when we recall, it will be a discrete, bounded engram and has a well-defined categorical boundary.

As far as system activity is concerned, I propose that categorisation is a manifestation of stability. The interaction between a cognitive agent and a particular object (book) has produced stable representations of the book which are invariant under most circumstances. I would like to propose that this pattern of invariants with regard to neural architecture could also be viewed as constituting a memory trace, an engram. All continuous experiences of particular objects (Harnad 1982) will be reduced to a memory trace that embodies necessary and sufficient features to be known as a category. Many researchers have pointed out that a category is a type not a collection of instances because what constitutes a memory trace would be patterns of invariants arising out of all continuous experiences.

It is suggested that this pattern of invariance arising out of system activity that undergoes stabilization of features when a cognitive agent interacts with an object is metaphorical in nature. Usually, we would relate categorisation to features of things instead of highlighting the different domains these things come from (we do not usually categorise chairs with desks, although we could). But features and domains are interconnected hence we could describe the metaphorical nature of categorisation based on objects in different domains. It is only when objects which exist initially in different domains are found to show similar features that they would be put under the same category. Thus, Gentner (1989, cited in Choe 2002) wrote that the concept of metaphor is the underlying, all embracing process that creates similarity when different domains are mapped onto each other. A penguin lives in a different domain than a robin but using similarity of features we could consider whether they would fall under the same category. Lakoff (1990) informed that Dyirbal, the Australian aboriginal language, categorises women, fire, and dangerous things of different domains under the same category called balan. Therefore, categorisation is metaphorical in nature where features which are considered similar are grouped under the same category. I think these examples show that describing categorization in terms of patterns of invariance would allow us to view categorisation as metaphorical. If Lakoff has described categorisation as basic to our thought, the system activity that gives rise to it metaphor, may be looked, artifactually, as a mechanism.

Therefore, the most basic of our thought is metaphorical in nature which helps us to recognize objects in our environment and to form decisions about them. This is the role that metaphor does as a system activity at the most basic level. Assuming that implicit thought processes generate explicit representations in the form of design ideas that we see in architecture students’ model making, it is my contention that if we look at students’ interpretation of their experience reflected in their model making, the first role of metaphor involving implicit categorising can be seen.

Their representation of water and fish scales to refer to sea and fish are acts of categorisation. They are merely performing acts of categorisation which is a very basic and dominant system activity in transferring their experience of sea and fish to their model. The elements of water and fish scales that are used to refer to objects in their environment are seen to fall under the same category of these objects of experience. ‘Classical categorisation’ is based on shared necessary and sufficient features. The features of the elements (water and fish scales) are shared with the objects of experience (sea and fish). Since categorization is a very basic system activity, students will perform this activity first before they are able to proceed to a different level of system activity which I shall argue later is manifested by the second role of metaphor.

2.2 Second Role of Metaphor

How could we understand the second role of metaphor from the interactivist framework? Again I shall pitch this understanding to Bickhard’s model of representationality. I would like to discuss this with reference to Gentner and Markman’s (1997) writing on structural alignment involving Kepler’s analogical reasoning between motive power that moves the planet around the sun and light (p.45). I suggest that this can be looked as a specialized system activity that allows learning to take place. This is level 8 in Bickhard’s levels of representationality. In this level, learning “…changes the nature of optimal organization within that interactive system” (1999: p.10). “It does involve prior knowledge, at least in a heuristic form – knowledge of sorts of problems associated with sorts of likely solutions” (2001b: p.2). Bickhard characterises learning as problem solving based on similarities between the new problems with new solutions to be found and old problems and old solutions of what worked in the past. He emphasizes the relationships of similarities when comparing between new problems and solutions and their corresponding old problems and solutions (2001b: p. 17). We do not have to go into detail to understand that heuristic problem solving in learning is also metaphorical. The key terms he uses i.e. relationships of similarities should remind us of metaphor. I would like to suggest, however, that the comparison to be made between ‘old knowledge’ and ‘new knowledge’ is not that of simple mapping of attributes that we looked at in categorisation but that of causal relations between them. Similarity involving causal relations requires that the structures which made up the things compared be highlighted instead of object attributes. As I have said earlier, this is important for learning and shall now argue for this using Kepler’s analogical example Gentner and Markman (1997: p. 46) provided.

In trying to understand what motive power drives the planets to move on a fixed orbit round the sun, Kepler used the analogy between light and the motive power. Describing this analogy Gentner and Markman (1997) wrote:

If light can travel undetectably on its way between the source and destination, yet illuminate its destination, then so too could the motive force be undetectable on its way from sun to planet, yet affect the planet’s motion once it arrives at the planet. But Kepler was not content with a mere proof of possibility. He pushed the analogy further. He used it to state why the motive power diminishes with distance: Just as the light from a lamp shines brighter on near objects than on further ones, so it is with the sun’s motive power, and for the same reason: The motive power (like the light) is not lost as it disperses but is spread out over a greater area.(p.46)

Kepler understood the old problem (illumination) and its solution (light, nearer brighter than farther) and applied this to new problem (movement of planet) and ‘found’ its new solution (motive power, nearer stronger than farther). Kepler was looking for a solution to a new problem he was facing. He was familiar with the light and its causal relations. These relationships between light and its illumination on objects were used by Kepler to provide causal relations between the sun and the revolving planets and ‘explain’ how the motive power is able to cause the planets to move and the differences in the planets’ motions. Kepler wasn’t comparing the similarities between the attributes of light and the planets to learn about the motive power but structural relations between the source (light) and the target (motive power) domains he was studying. It is when he was able to discover some kind of structural alignment between the source and target domains that the problem was solved and learning took place.

The consideration on similarity of structure between the source and target domains clearly differentiates this kind of metaphorical system activity from categorisation which operates on similarity of attributes. From the interactivist framework, Gentner and Markman’s (1997) structure alignment comparison is to be understood as an implicit system activity. Therefore both, mapping of attributes and structure alignment that they propose to account for cognitive processing are still private to the cognitive agent in nature. For real insights where scientific ‘creation’ (in the sense of Indurkhya’s) can take place, structure relations comparison induced by reality outside the private individual is required making scientific ‘creation’ public in nature. This will then give rise to the social structure of science (see Hull 2001) that scientists observed.

This role of metaphor is felt to be a higher level stage for novice designers. I would like to suggest that students go through the basic stage of categorization first before moving on to this higher level. As mentioned above, categorisation is involved in recognition, hence students must be able to recognize what they are dealing with before being able to notice or manipulate object properties related to causal relations in the sense of structure mapping Gentner and Markman discussed with regard to Kepler’s discovery. ‘Movement’, ‘change of surface from hard to soft’, ‘fluidity’ are ideas or properties connected to relationships between objects experienced and their environment.

2.3 Third Role of Metaphor

Indurkhya has written persuasively on the importance of similarity-creating metaphors (1998, 1999). In a nutshell, the role of metaphors is not just in “…pointing the existing similarities between …objects or situation” (p.45) but “…under certain conditions new perceptual and conceptual features can be created” (p.44) by metaphor.

Could I be so bold to state that similarity-creating metaphor that uses structure in the source domain to force a new way of looking at the target domain’s structure (but still preserving its old structure as suggested by Indurkhya (1998)) is a reality-induced phenomenon that implicit representational systems are not capable to process when shut off from reality. This statement seems contradictory since we know that the consciousness of reality-induced phenomena have to be stated by the individual, private cognitive agent who somehow then must have the internal capability to realize these structures in the first place. Otherwise, these structure cannot be stated. To overcome this paradox, I suggest that we look at implicit representational systems as only having the potential to be induced by the structure creating metaphors which reside in reality- hence in the public realm.

I argue that this public aspect of structure creating metaphor is an important trait that requires serious consideration. The history of science shows that public validation of scientific theories (by a community of scientists) has been instrumental in helping science progress in spite of the assertions of individual scientists.

Although knowing starts off as an individual experience, this experience does not guarantee scientific validity. Validation is to proceed by a community of experts. Therefore, having an individual experience of a phenomenon would just stay as internal knowledge regardless of its truth value as far as science is concern. Einstein had an individual experience of Theory of Relativity but as long as it stayed as internal knowledge unvalidated by the community of experts, it would not have been science. Therefore, this individual experience would not be a guarantee that knowledge out of individual experience complies with what scientific facts are which are about reality that resides outside the individual mind. The structure of scientific knowledge lies external to the cognitive agent. Scientists have noted this. Campbell, for example, once wrote that the search for scientific validity lies in the social structure of science (1997: p.13, cited in Hull 2001: p.158). Views proposed by scientists have to be tested and the tests be taken seriously (cited in Hull 2001: p. 165).

Maybe a reason why science requires validation external to the individual scientist is because it is impossible for an individual scientist to test all the proposals since scientific laws are spatiotemporally unrestricted. To prove that a law holds, this law must also be applicable elsewhere in the universe. But I like to propose another reason why the social structure in science in the form external validation with respect to metaphor is necessary- that is the knowledge of the structure of our universe which science ultimately deals with can only come about from the interaction between our mind and reality. This has been stated before at least by Indurkhya (1998, 1999). Eventually what science will have to uncover is the coherent structure of the whole universe. In other words, scientists will have to discover the complete constitution of the universe and how the elements that form this constitution are related causally. The mind is a specialized system activity that evolved to seek this structure. This structure does not reside in the mind otherwise, learning would not take place whereas we know that learning takes place all the time. However as stated by Indurkhya (ibid.), this structure can only be revealed from the interaction of the mind with reality. Thus, the mind is an all-encompassing metaphorical mechanism that slowly reveals this structure.

Since the mind has to interact with reality for the structure of the universe to be apparent, therefore it must then contain part of the mechanism that allows the whole structure that science seeks. I suggest that this third role of metaphor rides piggy-back on the second role using the similarity of causal relations seeking system activity. I would like to discuss further with some examples from Gentner and Markman (1997) in arguing for this conclusion.

In their paper titled Structure Mapping in analogy and Similarity, Gentner and Markman (1997) considered that using just the similarity-based account of comparison relying on common features, one is not able to distinguish between bats and birds because they have similar perceptual and behavioural characteristics. Hence, in their words, “…similarity’s role in categorization has been challenged” (p.54). I would like to argue that bats and birds can still be distinguished using similarity-based comparison of features. Someone who is familiar with both domains of birds and bats will be able to spot non-obvious differences, such as giving birth to live young and milk suckling. I consider these characteristics as features since they are patterns of invariance spelt out in the interactivist framework. Using bats and birds as an example to discredit similarity-based account does not quite work. A better example to argue against this account will be the ring species.

The ring species refer to populations of species that are found in a ring around the globe where interbreeding is possible if these populations are next to each other on this ring. However, one finds that at one point of this ring two species that happen to be next to each other do not interbreed hence would be considered as separate species. The gull species are such an example. In Britain, they are known as herring gulls which are white. Going eastward from Britain, white-coloured gulls are also found in Siberia. They are able to interbreed with herring gulls of Britain. As we go further east, in America, the gulls have black specks on them but are still able to interbreed with their Siberian counterparts. Further eastward of America the gulls have become black and hence called black-backed gulls. These gulls are not able to interbreed with the herring gulls of Britain.

The ring species should pose a more difficult problem for the similarity-based account to distinguish objects under separate categories based on features alone. This is because as gull species the birds would have a lot of features in common. The only way it seems to distinguish them would be to incorporate theory-based accounts proposed by Gentner and Markman (1997). This is because the theory used in biology to distinguish separate species would be the ‘biological species concept’ of non-interbreeding. Although this idea has been controversial, Mayr argued that this is still the most practical way to distinguish species (1988: p. 318-319). The theory based on the ‘biological species concept’ of non-interbreeding is used to distinguish the two gull species we referred to above putting them under separate categories.

The significance of the above explanation that I would like to assert is that the natural tendency for the mind to think in terms of classical categorization based on features would not provide a more coherent understanding of phenomena. However, the theory-based account of non-interbreeding is a structural consideration involving relationships between various objects of study that are found in reality outside the individual mind of the scientist. Indurkhya (1997: p. 11) in commenting Rosch’s idea of categorization said that “ …Reality, as it presents itself to us in the form of sensory stimuli that form the raw material for conceptualization and categorization, is already structured in fashion.” Hence, if reality is already structured, causal relations which structures are based on could only be prompted from reality not solely from the individual mind of the scientist.

Therefore, for the idea of non-interbreeding to be a discriminating concept, scientists have to work with various species that exist and study the ‘causal relations’ among them. Mayr (1998: p.140), in differentiating between the species taxon and higher taxa wrote “An essentialist (typological) definition is satisfactory and sufficient at the level of the higher taxa. It is, however, irrelevant and misleading to define species in an essentialistic way because the species is not defined by intrinsic, but relational properties.” [original emphasis] Explaining what he means by the terms relational properties, Mayr (ibid.) went on to say that “The word ‘species” likewise designates such a relational property. A population is a species with respect to all other populations with which it exhibits the relationship of reproductive isolation- - non-interbreeding.”

Embarking on the discovery the non-interbreeding definition of the species concept is a structural quest. Scientists need to have some idea how various populations of species are causally related to each other. This shows the importance of structural understanding of phenomena compared to understanding based on stable features. It is easier to observe featural similarities. Gentner and Markman reported that children’s development contains a lot of such ‘insights’ (1997: p.48). Hence the creation of structural relations by the mind which are then imposed on the environment is the most important way metaphor affects our understanding of the universe. This somehow can only take place if our mind has the necessary mechanism in place. I have proposed that the second role of metaphor which relies on similarity-based structural alignment as providing the basic system activity for similarity-creating metaphor to ride piggy-back on it.

This third role of metaphor is the most sophisticated of system activity. It is involved in the creation of new ideas and hence a new way of looking at phenomena. I shall not discuss this role further in this paper since my objective was only to understand why my students of architecture do the things they do which could be explained using the first two roles of metaphor. However, I believe this last role is even more important and shall be my research interest in the future.

3.0 Conclusion

The thinking processes whether of symbolic or non-symbolic representations that are depicted by these architecture students, it would be argued, arise out of the interpretive mechanism of metaphor. It is felt that this metaphorical mechanism is important in differentiating between novice designers and those who are more experienced. Designs that display non-symbolic interactive interpretation do not always mean that they are successful. The factors contributing to the success of a design are various. However, it is the belief of the researcher that non-symbolic interactive interpretation in design is displayed by mature designers hence, students of design should seek to achieve this mode of design thinking.


References

Bickhard, M. H. (1998a). Representational Content in Humans and Machines [online]. Available
at <URL: http://www.lehigh.edu/~mhb0/repcon.html> [ Accessed October 27th , 1998].

Bickhard, M. H. (1998b). Interaction and Representation [online]. Available at <URL:
http://www.lehigh.edu/~mhb0/interrepascii> [ Accessed November 1st 1998].

Bickhard, M. H. (1999). Levels of Representationality [online]. Available at <URL:
http://www.lehigh.edu/~mhb0/levels.html> [Accessed September 17th 1999].

Bickhard, M. H. (2001a). Critical Principles; On the Negative Side of Rationality [online].
Available at <URL: http://www.lehigh.edu/~mhb0/critical.html> [Accessed April 28th
2001].

Bickhard, M. H. (2001b). Topologies of Learning and Development [online]. Available at
<URL: http://www.lehigh.edu/~mhb0/topology.html> [Accessed April 28th 2001].

Black, M. (1962). Models and Metaphors. Ithaca, NY: Cornell University Press.

Choe, Y. (2002). Active Representations: A Primitive for Analogical Computing in the Brain.
American Association for Artificial Intelligence.

Gentner, D. (1989). The Mechanisms of Analogical Learning. In Vosniadou, S., and Ortony, A.,
(eds.), Similarity and Analogical Reasoning, 199-241. New York, NY: Academic Press.

Gentner, D., and Markman, A. B. (1997). Structure Mapping in Analogy and Similarity. American
Psychologist, 52(1), 45-56.

Harnad, S. (1982). Metaphor and Mental Duality [online]. Available at <URL:
ftp://ftp.princeton.edu/pub/harnad/Harnad/HTML/HTML/harnad82.metaphor.html>
[Accessed November 18th 1998].

Hull, D. L. (2001). In Search of Epistemological warrant. In C. Heyes & D. L. Hull (eds.) Selection
Theory and Social Construction: The Evolutionary Naturalistic Epistemology of Donald T. Campbell, 155-
167.

Indurkhya, B. (1992). Metaphor and Cognition: An Interactionist Approach. Dordrecht: Kluwer
Academic Publishers.

Indurkhya, B. (1994). “Metaphor as Change of Representation: An Interaction Theory of
Cognition and Metaphor”. In J. Hintikka (ed.) Aspects of Metaphor, Kluwer Academic
Publishers, Dordrecht, The Netherlands, pp.95-150.

Indurkhya, B. (1997). Metaphor as Change of Representation: an artificial intelligence
perspective. Journal of Experimental and Theoretical Artificial Intelligence. 9, 1-36.

Indurkhya, B. (1998). On Creation of Features and Change of Representation. Cognitive Studies,
Bulletin of the Japanese Cognitive Science Society, 5(2), 43-56.

Indurkhya, B. (1999). Creativity of Metaphor in Perceptual Symbol Systems. Behavioural and Brain
Sciences 22(4), 621-622.

Lakoff, G. & Johnson, M. (1980). Metaphors We Live By. Chicago: University Chicago Press.

Lakoff, G. (1990). Women, Fire and Dangerous Things: What Categories Reveal about the Mind. Chicago:
University Chicago Press.

MacCormac, E. R. (1985). A Cognitive Theory of Metaphor. Cambridge, MA: MIT Press.

Mayr, E. (1988). Toward a New Philosophy of Biology: Observation of an Evolutionist. Cambridge,
Massachusetts: Harvard University Press.

Mayr, E. (1998). Species Concept and their Application. In M. Ruse (Ed.) Philosophy of Biology.
New York: Prometheus Books, 136-145.

Ortony, A. (1979). Metaphor and Thought. New York: Cambridge University Press.

Richards, I. A. (1936). The Philosophy of Rhetoric. London: Oxford University Press.

BACK TO ISI 2003 HOME PAGE

ISI 2003 Home Page