The weekly column

Article 101, June 2002

The Case for Metaphoric Intelligence: A Reply

By Dr Jeannette Littlemore

In a recent contribution to Humanizing Language Teaching (2001a), I made a case for their being a 'Metaphoric Intelligence' (an ability to comprehend a produce novel metaphors) and suggested that this might bring a number of benefits to the language learning process. I also put forward the idea that metaphoric intelligence constitutes a ninth 'intelligence type' to be added to the eight that have already been suggested by Gardner in his theory of Multiple Intelligences (1993; 1999). Recently, in this column, (April, 2002 - ) Michael Berman contested this idea, asserting that metaphoric intelligence meets none of Gardner's (1993) eight criteria for the existence of an intelligence type, and that it therefore should not be included in the Multiple Intelligence theory. Berman went on to suggest that metaphoric intelligence is best perceived as an aspect of spiritual intelligence (SQ).

In this article, I would like to take issue with both of Berman's claims. Firstly, I provide theoretical and empirical evidence showing that metaphoric intelligence does indeed meet Gardner's (1993) eight criteria for the existence of an intelligence type. Secondly, I explain why metaphoric intelligence is not simply an aspect of spiritual intelligence.

Gardner's first criterion for an intelligence type is that it should be associated with a specific area of the brain, and that damage to that area of the brain adversely affects that type of intelligence. Metaphoric intelligence appears to meet this criterion. Using a PET scan [1], Bottini et. al. (1994), were able to locate the exact parts of the brain that are activated during metaphoric processing. They read to their subjects a number of plausible metaphors and asked them to decide whether or not they found them meaningful. The subjects were also read a number of literal sentences and asked to decide whether or not they were meaningful. A PET scan was taken of the subjects' brains while they were making each decision. Bottini et. al. found that the processing of metaphors gave rise to activation in a number of areas of the brain which were not activated during the processing of literal sentences. These significant activations were observed in the right hemisphere frontal lobe, the right hemisphere temporal lobe, and the right hemisphere precunius. There is also plenty of evidence showing that damage to the right hemisphere of the brain can lead to a severe disfunctioning in metaphor interpretation abilities (see, for example, Hier and Kaplan, 1980; Wapner, et al., 1981; Brownell et al., 1990).

Another of Gardner's criteria is that in order for an intelligence type to exist, it should have an identifiable set of core operations. I have argued elsewhere (1998; 2001a; 2001b) that the two main processes likely to be involved are associative fluency and analogical reasoning, and that these processes are aided by mental imagery. Empirical support for this view comes from extensive investigations that have been made into the psychological correlates of metaphor comprehension and production. For example, in order to trace some of the psychological processes underlying different kinds of figurative competence, Pollio and Smith (1980) carried out a detailed factor analysis of a number of tests of figurative competence. They were able to identify an 'associative fluency' factor, and a 'syllogisms' factor.

Associative fluency refers to the ability to make a wide range of connections when presented with a given stimulus. Support for the relevance of associative fluency to metaphoric fluency can be found in Miller's (1987) description of divergent search strategies for retrieval from memory. He claims that divergent searches are broad and associational rather than logical, and that they rely on vague search criteria. When asked to find multiple interpretations for a metaphor, subjects need to search the semic webs (i.e. the networks of associations) surrounding each of the two parts of the metaphor in order to find areas of overlap. Individuals with a divergent search strategy will make a broader search, including meanings that are not included by those with a convergent search strategy. The two semic webs will therefore be larger and there will be more potential for overlap between them. This would lead one to expect divergers to provide more interpretations than convergers; in other words they are more likely to perform well on tests of metaphoric fluency.

Support for a relationship between associative fluency and metaphoric fluency also comes from findings made by Carroll (1993). An examination of a number of tests designed to measure the ability to produce large numbers of "related ideas" enabled him to identify a highly specific "associational fluency" factor. In order to load on this factor, the measures must apparently require that:

a series of associations are to be given, and the score is the number of associations produced (written) in a given time. (ibid.;414)

Carroll found that a test in which subjects were asked to think of different ways of completing unfinished similes loaded on this factor.

The characteristic common to all the tests that load on Pollio and Smiths' 'syllogisms' factor is that they all rely on some form of comparison. This implies that these tests depend upon the psychological process of analogical reasoning. Analogical reasoning is a process whereby partial similarities are observed between concepts so that the characteristics of one of the concepts can be used to shed light on the other. It is of particular interest to psychologists as it is a device often used by people to help them understand new concepts. According to Gentner (1983) the process of analogical reasoning can be decomposed into a number of subprocesses:

- accessing the source
- performing the mapping between base and target
- evaluating the match
- storing inferences in the target
- extracting the commonalities

This process has a great deal in common with metaphoric processing, which also involves the capacity to perceive partial similarity between different domains. The role of analogy in metaphoric processing has been outlined by Paivio and Walsh (1993) who claim the capacity to perceive partial similarity between apparently dissimilar domains is central to all kinds of metaphoric processing. They claim that the basis of similarity may lie in shared attributes, or that it may involve relational similarity. It is noteworthy that these are also the two main bases for analogical reasoning. This implies that analogical reasoning and metaphoric processing are closely related.

The use of mental imagery is likely to help individuals to engage in analogical reasoning. Paivio and Walsh (1993) argue that different pieces of information that are represented visually can be recalled simultaneously, whereas different pieces of information which are represented verbally can only be recalled sequentially. Harris et al. (1980) examined subjects' use of imagery in encoding metaphors in comparison to non-metaphors. They found that subjects typically reported vivid images in reaction to metaphors. They used images significantly more frequently to encode metaphorical sentences than non-metaphorical sentences. Interactive images were also present, which, in Harris et al's words:

Were frequently highly creative, constructed, literally anomalous 'surrealistic' images involving both the topic and vehicle fused in dynamic interaction (op. cit.;178)

By forming interactive images, individuals are better able to see relationships between both parts of the metaphor. Thus we can say that at least three processes are likely to underlie metaphoric intelligence, namely associative fluency, analogical reasoning, and image formation.

Gardner's next criterion for an intelligence type is that it should have a distinctive developmental history along with a definable set of end-state expert performances. Research on children has suggested that both metaphoric production and metaphoric comprehension are skills that increase with age. Much of this research was carried out by Gardner himself. Gardner et. al. (1974) carried out an experiment into childrens' capacities to both create and appreciate metaphors. In the first part of their experiment, the students were instructed to think of appropriate metaphorical endings for a list of twelve very short stories. In the second part the students were told to choose the "best" metaphor from a list of four metaphors containing a novel metaphor, a literal comparison, a conventional metaphor and an inappropriate metaphor. In each case, they were told that marks would be gained for novelty.

In one of the questions, the students were instructed to think of an appropriate metaphor to complete the following sentence:

Peter came sneaking into the room. It was clear that he had experienced something unusual. His voice was...

They then had to choose their preferred ending from the following four endings:

LITERAL: the quietest sound we've heard
CONVENTIONAL: like a mouse sitting in the room
NOVEL: like a dawn in a ghost town.
INAPPROPRIATE: like a family going on a trip.

Gardner et. al. observed a tendency, increasing with age, towards novel metaphoric production and preference. They also found that the highest percentages of novel metaphors were produced by the youngest subjects and the oldest subjects. Where these two groups differed was in the tendency of young subjects to produce metaphors which were highly original but inappropriate or nonsensical.

Support for the development of metaphoric intelligence has also been demonstrated by Kogan (1983) who administered the Metaphoric Triads Task (MTT) to children of different ages. The MTT is a non-linguistic test comprising 29 triads of pictures. Each triad offers three pairing possibilities. For example, one triad offers pictures of a toddler on the grass, a brightly coloured watering can, and a rosebud. The following pairing links are possible:

Toddler and watering can: The toddler can play with the watering can;
Watering can and rosebud: One can sprinkle water on the rosebud with the watering can;
Toddler and rosebud: Both are at an early stage of development

The third link is considered to be the metaphorical link. The children are asked to make as many pairings as possible between the pictures. As this test has no overt linguistic element it is hypothesized to be a test of pure cognitive development. Using this test, Kogan found a progressive increase in metaphoric comprehension across an age span extending from kindergarten to young adulthood.

We have thus seen that metaphoric intelligence does have a distinctive developmental history. As for the set of 'end-state performances', Gardner himself, when referring to metaphoric ability, talks of "the indisputable existence of a developed end-state" (1993;294). As evidence, he refers to the existence of skilled metaphorizers in fields such as poetry and art. The existence of such skilled individuals also provides support for Gardner's next criterion that there should be 'idiots savants', prodigies, and other exceptional individuals. It is easy to cite examples of prodigies and other exceptional individuals in the area of metaphoric intelligence (one only has to think, for example, of Shakespeare). There is less evidence for the existence of idiots savants (individuals who are excellent at metaphorizing but who lack other types of intelligence). However, this does not mean that they do not exist. It could simply mean that they are not easily identifiable.

Gardner also believes that different types of intelligence should be measurable through experimental psychological tasks, and that there should be support from psychometric findings. The best way of isolating metaphoric intelligence is through the Metaphoric Triads Test (MTT) described above. As the MTT is a purely pictorial test, it cannot be accused of measuring linguistic skills and is therefore the purest possible test of metaphoric intelligence. The fact that this test is able to measure metaphoric intelligence with reasonable reliability goes a long way towards meeting these two criteria.

Another of Gardner's criteria is that an intelligence should be susceptible to encoding in a symbol system. Metaphor itself is a highly symbolic activity. A 'symbol' is an object or device that is traditionally used to refer to something else. In their seminal work on conceptual metaphor, Lakoff and Johnson (1980) demonstrate how our entire conceptual system is symbolised through metaphor. For example, evidence for the conceptual metaphor AN ARGUMENT IS A CONTAINER is provided via expressions such as the argument doesn't have much content, the argument won't hold water, I'm tired of empty arguments, and so on.

This leaves Gardner's final criterion: an evolutionary history and evolutionary plausibility. Gardner warns that this is an area where firm facts are particularly elusive and speculation is tempting. Whilst keeping this warning in mind, one could take Lakoff and Johnson's argument that metaphor forms the basis of all abstract thought and conclude that our ability to metaphorize has played a key evolutionary role in distinguishing humans from other species, allowing us to survive on our skills of abstraction, rather than our physical capacities. This argument lends a degree of evolutionary plausibility to the metaphoric intelligence construct.

We have therefore seen that metaphoric intelligence meets all of Gardner's criteria and that it might therefore be considered to be a type of intelligence. I would now like to discuss Berman's argument that metaphoric intelligence is simply an aspect of SQ. According to Gardner (1999), SQ comprises 'a desire to know about cosmic experiences that are not easily apprehended in a material sense' (ibid.;55), 'an ability to achieve a trance-like state' (ibid.;55), and 'an ability to have a spiritual affect on others' (ibid.;57). In my mind it is difficult to see how any of these are relevant to the concept of metaphorical intelligence. Furthermore, according to Berman, SQ involves, among other things, an ability to be field-independent. It has been empirically demonstrated that metaphoric intelligence is significantly, negatively related to field-independence and that it is more likely to be related to holistic, synthetic processing (Johnson and Rosano, 1993; Littlemore, 2001b). Although metaphoric intelligence may have points in common with SQ (such as an ability to see a connection between diverse phenomena), theoretical and empirical evidence suggest that it is more appropriately viewed as a distinct type of intelligence. Whether we view metaphoric intelligence as an acquired skill, or as a distinct intelligence, the fact remains that it is likely to have a number of useful applications in language learning. It should enrich language production and facilitate the comprehension of metaphoric expressions. It is therefore likely to contribute positively to an overall level of communicative competence.


Berman, M. (2002). The case for metaphoric intelligence. ELT Newsletter, .
Bottini, G., Corcoran, R., Sterzi, R. and Schenone, P. (1994). The role of the right hemisphere in the interpretation of figurative aspects of language. A positron emission tomography activation study. Brain, 117 (2): 1241-1258.
Brownell, H., Simpson, T., Bihrle, A., Potter, H. and Gardner, H. (1990). Appeciation of metaphoric alternative word meanings by left and right brain-damaged patients. Neuropsychologia, 28 (4): 375-383.
Carroll, J.B. (1993). Human Cognitive Abilities. A Survey of Factor-Analytic Studies. Cambridge: Cambridge University Press.
Gardner, H. (1993). Frames of Mind (2nd Ed.). London: Fontana Press.
Gardner, H. (1999). Intelligence Reframed. New York: Basic Books.
Gardner, H., Kircher, M., Winner, E. and Perkins, D. (1974). Childrens' metaphoric productions and preferences. Journal of Child Language, 2: 125-141.
Gentner, D. (1983). Structure-mapping: A theoretical framework for analogy. Cognitive Science, 7: 155-170.
Harris R.J., Lahey, M.A. and Marsalek, F. (1980). Metaphors and images: Rating, reporting and remembering. In R.P. Honeck and R.P. Hoffman (Eds.) Cognition and Figurative Language, New Jersey: Erlbaum Press.
Hier, D.B. and Kaplan, J. (1980). Verbal comprehension deficits after right hemisphere damage. Applied Psycholinguistics, 279-294.
Johnson, J. and Rosano, T. (1993). Relation of cognitive style to metaphor interpretation and second language proficiency. Applied Psycholinguistics, 14: 159-175.
Kogan, N. (1983). Stylistic variation in childhood and adolescence: Creativity, metaphor and cognitive styles. In J.H. Flavell and E.M. Markman (Eds.) A Handbook of Child Psychology. New York: John Wiley and Sons.
Lakoff, G. & Johnson, M. (1980). Metaphors We Live By. Chicago: Chicago University Press.
Littlemore, J. (1998). "Individual differences in second language learning: Towards an identification of the strategy preferences and language learning strengths of students with a holistic and/or imager cognitive style". Unpublished PhD Dissertation. Thames Valley University.
Littlemore, J. (2001a) Metaphoric intelligence and foreign language learning. Humanizing Language Teaching 3, 2.
Littlemore, J. (2001b). "Metaphoric competence: A language learning strength of students with a holistic cognitive style?" TESOL Quarterly, 35 (3): 459-491.
Miller, A. (1987). Cognitive styles: An integrated model. Educational Psychology, 7 (4): 251-268.
Paivio, A. and Walsh, M. (1993). Psychological processes in metaphor comprehension and memory. In A. Ortony (Ed.) Metaphor and Thought Cambridge: Cambridge University Press.
Pollio, H. and Smith, M. (1980). Metahoric competence and complex human problem solving. In R.P. Honeck and R.P. Hoffman (Eds.) Cognition and Figurative Language, New Jersey: Erlbaum Press.
Reichmann, P. and Coste, E. (1980). Mental imagery and the comprehension of figurative language. Is there a relationship? In R.P. Honeck and R.P. Hoffman (Eds.) Cognition and Figurative Language, New Jersey: Erlbaum Press.
Wapner, W., Hamby, S., and Gardner, H. (1981). The role of the right hemisphere in the apprehension of complex linguistic materials. Brain and Language, 14: 15-33.

[1] By means of a technique known as positron emission tomography (PET), researchers can study visual displays of the locations in the brain that are active during a variety of tasks, including those tasks involved in language use. This experimental technique depends on the fact t
hat more blood is delivered to the parts of the brain that are active than to the parts that are relatively inactive. A slightly radioactive substance that can be picked up by detectors surrounding the head is introduced into the blood. The subject is asked to carry out various tasks and the resulting brain activity is mapped in three dimensions by the PET scanner.

About the Author

Jeannette Littlemore ( is a lecturer in English for Academic Purposes at the University of Birmingham, UK. She has a PhD in Individual Differences in Second Language Acquisition. She is currently carrying out research into metaphor and language teaching, and the relationship between cognitive style and second language learning. She has also taught English in Spain, Japan and Belgium.

Questions or comments about this week's article? Why not post them on our Discussion Forum