Consciousness, Literature and the Arts

Archive

Volume 2 Number 2, August 2007

_______________________________________________________________

Metaphor as Guided Experience:  Metaphors and Situation Models

By

Kimberly Mosher Lockwood

University of Dayton

But in fact there is no limit to what a metaphor calls to our attention, and much of what we are caused to notice is not propositional in character.  When we try to say what a metaphor ‘means’, we soon realize there is no end to what we want to mention.  If someone draws his finger along the coastline of a map, or mentions the beauty and deftness of a line in a Picasso etching, how many things are drawn to your attention?  You might list a great many, but you could not finish because the idea of finishing would have no clear application.  How many facts or propositions are conveyed by a photograph?  None, an infinity, or one great unstatable fact?  Bad question.  A picture is not worth a thousand words, or any other number.  Words are the wrong currency to exchange for a picture.

                                                                   -Donald Davidson^[1]

Metaphors seem to be able to affect us strongly.  Some theorists believe that they can reshape our cognitive structure and impact our behavior.  In an attempt to explain the power that metaphors have, I will propose a theory of metaphor as guided experience.^[2]  Metaphors are not interpreted as isolated speech acts but as part of a larger context.  I propose that this context and the metaphor itself cue hearers to construct a mental model called a situation model.  This situation model recreates a  state-of-affairs, much like a mental movie but more complex.  Experiencing this situation model provides a guided experience of both the larger context and the metaphor itself.  I will attempt to show that this experience explains how metaphors can influence a person’s cognitive processes.

          In order to create and explain this theory, I will begin with a brief discussion of Max Black’s influential theory of metaphor.  I will then present the work of George Lakoff and Mark Johnson, who offer a prominent theory of conceptual metaphors.  Their work helps us define what a metaphor’s cognitive effect really is, but it is weak in its explanation of how this effect is created.  I will offer an explanation to replace Lakoff and Johnson’s that is based on a theory of situation models and incorporates the work of Max Black.  Once the theory of metaphor is explained, I will apply it to metaphors found in Wordsworth’s “I Wandered Lonely as a Cloud.” I conclude with an explanation of how I believe this theory is an advancement of our understanding of metaphor.

I.  Max Black

Max Black is often credited with being the first analytic philosopher to give the problem of metaphor any serious analysis.  Black recognizes that metaphors pose problems that are not only pragmatic but also semantic.  A statement of his interaction theory appears in "Metaphor".^[3]  To summarize, Black argues that a metaphorical statement has two subjects: the principal and the subsidiary subject.  For example, in the metaphor, 'Man is a wolf,' the principal subject is 'man' and the subsidiary subject is 'wolf'.  The subjects are regarded as 'systems of things' rather than things.  "Wolf" does not just refer to wolves but to a system of commonplaces associated with wolves (e.g. being aggressive, fierce, or predatory).  In a metaphor, the commonplaces of the subsidiary subject are evoked, and implied assertions regarding the principal subject emerge.  Black explains:

The effect, then, of calling a man a 'wolf' is to evoke the wolf-system of related commonplaces.  If the man is a wolf, he preys upon other animals, is fierce... Each of these implied assertions has now to be made to fit the principal subject either in normal or abnormal senses.  A suitable hearer will be led by the wolf-system of implication about the principal subject.  But these implications will not be comprised of commonplaces normally implied by literal uses of 'man'.  The new implications must be determined by the pattern of implication associated with literal uses of 'wolf'.  Any human traits that can without undue strain be talked about in 'wolf-language' will be rendered prominent, and any that cannot be will be pushed into the background.  The wolf metaphor suppresses some details, emphasizes others-in short, organizes our view of man.^[4] 

          This theory has been criticized for failing to pick out the appropriate commonplaces for the interpretation of a metaphor.  There is the possibility that a subsidiary subject might have an endless list of commonplaces to be considered and that not all that may apply may be relevant.^[5]  Many commonplaces of 'wolf' can be literally applied to 'man', such as being a mammal, which are neither informative nor do they seem to actually be implied by the metaphor.  How, on Black's account, can these irrelevant commonplaces be eliminated from consideration?  Black attempts to remedy this problem in his essay, "More about Metaphor".^[6]   In this essay he explains that it is the interaction between the two subjects that determine which commonplaces are relevant.  Different principal subjects will pick out different commonplaces of the subsidiary subject.  Black explains by comparing the 'man is a wolf' metaphor to a 'wind as a wolf' metaphor found in the following sentence:

          In the deserted land, the wind, a lone wolf, howled in the night.

The commonplaces in the 'man is a wolf' metaphor (ferocity, being predatory, etc.) are no longer applicable to the principal subject.  The principal subject will determine which commonplaces are implied by the metaphor.

          Eva Feder Kittay argues that there are still two problems with this theory.^[7]  First, she argues that the above criticism has not been sufficiently answered.  The commonplaces of the wolf-system will apply to 'man' but will not be helpful in interpreting the metaphor.  For example, men and wolves are both living beings, carnivorous, and mammals, though these traits are not what the metaphor means to convey.  Black argues that interpretations will be limited by the significant commonplaces shown through the interaction of the principal and subsidiary subjects, but there is no way to determine which commonplaces are significant and which are not.  Secondly, she adds a new criticism that more complex or more creative metaphors do require "strain on language and thought."^[8]  Some examples that Kittay uses that demonstrate this point are the following metaphors:

The garden was a slum of bloom.^[9]

Then how should I begin to spit out all the butt-ends of my days and ways?^[10]

In metaphors such as these the principal subject rejects the application of the subsidiary subject's commonplaces and this is part of the effect of the metaphor. The commonplace of ‘slum’ are not easily applied to ‘garden’; they seem to be contrary to what we usually think of when we refer to a garden.

Black has successfully argued that the effect of a metaphor is at least partially dependent upon the literal meanings of the words involved; however, two important problems remain.  I suggest that these problems can be solved once we can determine:  1) how we eliminate trivial similarities and associations of metaphorical statements, and 2) how it must determine what metaphors’ cognitive effect can consist of and it must also attempt to explain what content a metaphor can have.  My main interest lies in claims that metaphors can give us insight, change how we think and feel about something, and even change our cognitive structure.^[11]  My goal is to discover how metaphors can have a cognitive effect on their audience and to define what the nature of this effect is.  I will start this project by putting the work of George Lakoff and Mark Johnson^[12] forth as a fruitful basis for exploration.  I see my contribution to this discussion largely as offering a plausible explanation for some of the evidence offered by Lakoff and Johnson and as a way of filling in some gaps I find in their theory of conceptual metaphors.  I will argue that while Lakoff and Johnson can offer a view of what the cognitive effects of metaphor are, a theory of metaphors and situation models will answer the question of how this effect is achieved.

What is the cognitive effect of a metaphor?

Donald Davidson and Lakoff and Johnson have offered us ideas about an array of cognitive effects that metaphors may have from the ability to direct our attention similarly to that of a gesture (i.e., pointing to an area on a map), to the capability of changing the way we understand things which, in turn, results in behavioral changes.  I will argue that these effects may be explained only by a theory that can explain how literal language can affect us.  This is not a theory that explains how we understand or interpret literal language.  I begin by assuming that audiences possess some prior linguistic knowledge.  That being the case, what is the difference between understanding language and language affecting one’s cognition?  I will argue that language cues or stimulates the audience to create a situation model, a model of the state-of-affairs being described that is much like a movie of the events described.  In doing this, we are allowed to experience this state-of-affairs as guided by the description given.  Such a guided experience can have effects on our cognition in a similar way to that of experience itself.  In the case of metaphor, this experience is essential to its cognitive effect.  The situation model constructed provides the context in which we can interpret the metaphor.  This idea will be illustrated throughout the chapter.  The first step in this explanation is determining what the possible effect of metaphor might be.  For this, I will turn to the work of Lakoff and Johnson.

          In Metaphors We Live By, Lakoff and Johnson look at language for evidence of our metaphorically structured conceptual systems. They believe that because language is a product of our conceptual system it is an “important source of evidence for what that system is like.”^[13]  While I am not convinced that our conceptual system is metaphorically structured, I am interested in reinterpreting what they claim as evidence for their view. 

          To support their theory, Lakoff and Johnson give linguistic examples that they believe demonstrate that we really think metaphorically about many elements of everyday life. Further, they argue that this way of thinking influences our everyday actions. Lakoff and Johnson take these actions and everyday behaviors as further evidence of what our conceptual system is like, and they offer several examples to help make their point. Consider the metaphor ARGUMENT IS WAR, which they consider a conceptual metaphor. They point out several everyday expressions that are evidence of this conceptual metaphor, and they believe that how we conceptualize our actions before, during, and after an argument are primarily the results (or further evidence) of this conceptual metaphor. All of this attests to how systematically the conceptual metaphor influences our thinking, resulting in certain actions and behaviors. They explain:

It is important to see that we don’t just talk about arguments in terms of war. We can actually win or lose arguments. We see the person we are arguing with as an

opponent. We attack his positions and define our own. We gain or lose ground. We plan and use strategies. If we find a position indefensible, we can abandon it and take a new line of attack. Many of the things we do in arguing are partially structured by the concept of war. Though there is no physical battle, there is verbal battle, and the structure of an argument - attack, defense, counterattack, etc. - reflects this. It is in this sense that the ARGUMENT OF WAR metaphor is one that we live by in this culture; it structures what we perform in arguing.^[14]

The evidence that we speak and behave in this way is all around us.  Metaphors seem to be able to structure and change our knowledge of the world to a certain extent.  Lakoff and Johnson explain that we can only understand this if we have some idea of language and its role in knowledge in the first place.  In order to answer this question Lakoff and Johnson use the work of researcher Eleanor Rosch to explain what a concept is and use her theory as their own.  Rosch’s theory is that our concepts are categories formed through the sorting of things and ideas that have certain commonalities with a category’s prototypes.  She calls this categorization in terms of family resemblance.  Lakoff and Johnson use the example of the category “birds” to explain this idea.  The category is defined by its prototypes; for birds these are robins, sparrows, etc.. These birds have most of the family traits; being able to fly, singing, etc..  Other birds, like ostriches and penguins which are not prototypical birds, are recognized as birds by the traits they have in common with the prototypical birds, for example, being winged and two legged.  This is also the way Lakoff and Johnson believe metaphorical concepts are developed.                       

          There is a problem with Lakoff and Johnson’s acceptance of Rosch’s theory in that it is flawed as an explanation for how concepts are formed.  Rosch’s error is that she requires that a person already possess certain concepts in order to form concepts.  To recognize that an object has certain traits is to conceive of the object as having those traits, to apply a concept.  One cannot categorize a bird as a bird without applying the concept of “flies”.  In other words, Rosch is begging the question by using the idea of people having certain concepts to explain how concepts are formed.  I will return to this issue shortly.

          Lakoff and Johnson argue that conceptual metaphors are grounded in human experience.  Common experiences become the prototype experiences and families or concepts are formed when new experiences are sorted according to how similar the experience is to a prototypical experience.  They explain: “…What we are claiming about a (conceptual metaphor’s) grounding is that we typically understand the nonphysical in terms of the physical – that is, we conceptualize the less clearly delineated in terms of the more clearly delineated.”^[15]  Lakoff and Johnson offer basically two main types of conceptual metaphors grounded in the manner described, orientation metaphors and ontological metaphors.  Experiences of physical orientation, cultural beliefs, etc. become prototypes for metaphorical concepts.  Orientational metaphors are developed when a similarity is noticed between the new experience and a physical experience already understood.  Some examples of how this works offered by Lakoff and Johnson are

CONSCIOUS IS UP; UNCONSCIOUS IS DOWN

Get up, Wake up, I’m up already. …He dropped off to sleep. He’s under hypnosis.  He sank into a coma.

     Physical basis: Humans and most other mammals sleep lying down and stand up when they are awaken.^[16]

HAVING CONTROL AND FORCE IS UP; BEING SUBJECT TO CONTROL OR FORCE IS DOWN

I have control over her. …He is under my control. …He fell from power. …He is low man on the totem pole.

     Physical basis:  Physical size typically correlates with physical strength, and the victor in a fight is typically on top.^[17]

As we might come to understand new experiences in terms of physical orientations, we also may come to understand new experiences through their family resemblance to entities and substances that are familiar to us.  Lakoff and Johnson call this the forming of ontological metaphorical concepts.  One example of this is the INFLATION IS AN ENTITY metaphorical concept:

Inflation is lowering our standard of living. …inflation is backing us into a corner. …inflation makes me sick.

          In these cases, viewing inflation as an entity allows us to refer to it, quantify it, identify a particular aspect of it, see it as a cause, act with respect to it, and perhaps even believe that we understand it.^[18]

          Many ontological metaphors take the form of personification.  Explaining the nonhuman in terms of the human is probably so common because we have vast experience with human behavior – being human and being involved with other humans.  Lakoff and Johnson believe that metaphors are made directly from our experiences and, if this is true, it makes sense that our most intimate experiences would be the basis for most of our metaphors.

          The evidence that Lakoff and Johnson present is persuasive.  The metaphors they use to create their categories are so common that we have all probably encountered and used them at some point in time.  The main problem is that we are not told how these categories of metaphor can become so pervasive in the lives of individuals and of a culture.  What makes metaphors able to change the way we think and behave?  I believe that a theory of situation models can provide us with the answers to these questions.  Before I can build this theory, I first need to address the problem left by Lakoff and Johnson’s use of Rosch’s work as their basis for their theory.  Another way of developing knowledge is needed to understand how language can have cognitive effect.  Most of the categories of Lakoff and Johnson’s conceptual metaphors explain something abstract in terms of concrete experience.  How do we represent knowledge of the world we experience?  To answer this question, I will rely on a theory that attempts to explain that the brain represents knowledge as it is perceived in our actual experience.

Representing our perceptual experiences

An approach that may offer a better basis for this project is the theory of knowledge as perceptual symbol systems offered by Lawrence J. Barsalou.^[19]  If we make metaphors by understanding more abstract concepts in terms of more basic experiences, then it will be useful to understand how our more basic experiences might be represented in our cognitive system.  Barsalou relies on the fairly uncontroversial assumption that certain elements of our immediate experiences are extracted through selective attention and stored in long-term memory (LTM).  This process results in a collection of symbols, which are records of neural representations, and they are stored schematically or framed (i.e., organized by size, shape, color, etc).  Since we experience these elements as gestalts and represent them all at once, prior conceptual knowledge is not necessary.  Barsalou leaves the nature of these representations vague.  He believes that many theories of how the brain represents will fit with his overall ideas.  He describes these representations as a collection of brain states or other mental representations to be determined by future research.  The process of representation can have a related conscious experience (where we are actively trying to commit a thing to memory), but often we are not consciously focusing on learning something or memorizing when the representation is developed.  Findings reported in the literature on skill acquisition support the idea that at first we may have conscious experience that coincides with forming neural representations, but as we develop the skill unconscious mechanisms take control.^[20]  Learning to drive is a good example of this process.  When we first operate a vehicle, we actively think about reaching for and turning on turn signals, moving our foot to the brake when we see a red light, etc.  Eventually, however, these become automatic, unconscious activities.  Often we see a red light and simply respond without consciously thinking.  This example also illustrates how a perceptual cue can activate unconscious activity.  We see the light and react without thinking about that action specifically.

          It is necessary to notice some important implications of Barsalou’s ideas.  We can assume from the fact that our perceptual symbol systems are built from our personal experiences, that each person’s representations will differ to some degree.  We will have similarities based upon the similarities of our possible experiences.  We all experience the same physical world governed by certain laws and containing certain elements.  However, our representations will be idiosyncratic.  We will all create our representations from differing sets of experiences.  Further, we all have different sensory capacities and neural mechanics. 

          We will all develop automatic neural responses to certain cues and I will call these responses “dispositional representations”.^[21]   As we have sensory experiences we record them.  We represent our sensory experiences in the brain as potential neural firing patterns that can be recalled and reconstructed again and again.  These potential firing patterns are called “dispositional representations.”  It is in this way that we store and recall what we have experienced.  When cued by language (or any other cue), we reconstruct an image from past experiences to approximate the unfamiliar described situation.  This reconstruction is not necessarily something of which we are conscious.  For example, it is a dispositional response that underlies the behavior of depressing the brake when the cue ‘red traffic light’ is spotted.  Events that we have stored as firing patterns that tend to occur concurrently will come to be associated with one another.  If it is the case that I tend to experience the sensation of cold with the sight of snow, the two potential firing patterns will come to be associated or be cued to fire together.  When the sight of snow occurs, the sensation of cold will also be cued, and vice versa.  When dispositions are associated, this means that when one is cued, the other will likely also be cued.  These dispositional responses are often a product of our idiosyncratic representations and they are the building blocks of a situation model. Examples later in this article will illustrate this idea.

Situation models

I propose that language acts as a cue to the brain to represent the information given.  Language may act as a cue in the same way that coming to a red light makes us step on the brake without thinking about it.  I hypothesize that when we encounter language we automatically, and often unconsciously, represent the state-of-affairs described.  This representation is what I will call a situation model.  A situation model is much like a movie we create built out of our previously developed perceptual symbols.  This movie is multimodal; it may include visual, aural, and olfactory representations as well as their accompanying emotional and dispositional associations.  I speculate that the basic components of situation models include representations of entities (people, objects, ideas, aromas etc.), and properties of entities (color, emotions, goals, shape, etc.).  These components are structured in accordance to spatial-temporal frameworks (spatial locations, time frames), and relational information (spatial, temporal, causal, ownership, kinship, etc.).  Since situation models will vary from person to person due to individual idiosyncrasies, it will be difficult to define correct and incorrect Situation models very precisely.  However, I believe we can loosely define what constitutes a range of appropriate situation models in an intuitively appealing way.  Situation models are within an appropriate range when they depict plausible situations based upon the typical uses of the words that cue them.  Appropriate situation models must mirror the logic that governs the everyday world as we experience it, unless a description asks us to believe something contrary to our experience.  For example, a science fiction novel may ask us to imagine that teleportation is possible.  An inappropriate situation model flouts typical usage of the terms that cue it. Language acts as a set of processing instructions by which the comprehender constructs a mental representation of what is being described. 

          Before proceeding, it is necessary to acknowledge that there are other theories that regard some sort of mental simulation as the key to our understanding of narrative and that these simulations are responsible for our emotional responses to narrative and for narrative’s potential cognitive effect.^[22]  The first difference with this situation model theory is that it is not restricted to the understanding of narrative, but it is proposed as a response to all encounters with language from conversations with friends to reading novels.  Because many theories focus only on narrative, mental simulation is sometimes considered an act of the imagination, where imagining is a conscious cognitive process that requires one to put oneself in the situation of a fictional character and empathize with her/him in a particular circumstance.  In this case, a mental simulation is not based on empathy and it is not a conscious activity as imagination is sometimes regarded but the development of a situation model whose details are cued or caused automatically by everyday communication.  Situation models are to be understood as a part of everyday thought that can also be used as a tool in what is popularly considered imagining.   The important difference is that we typically think of imagining as something that one sets out to do and that is consciously experienced, whereas a situation model is to be understood as an automatic, involuntary response of which one may or may not be conscious. 

 Examples of how situation models are formed and what can be represented

In order to see what form a situation model would take, I will illustrate the idea through a few examples.  I will start with a simple sentence. 

(1) The wet dog lies in front of the fireplace.

The situation model for this sentence for most people would include a dog, lying in some position that is in front of a fireplace.  Notice that these things are not specific.  The dog does not have to be a German shepherd on his back in front of a brick fireplace with a glass door.  However, most people will have such specifics in their situation models, often without even noticing that they are filling in the blanks.  This is due to the fact that what is being cued is derived from a specific person’s set of idiosyncratic, dispositional representations.  For example, I picture a black lab/chow mix curled up on a red plaid dog bed in front of a brick fireplace.  This is my mutt, Abby.  I have additional modes to my representation because Abby happens to emit a foul odor when wet, despite which I still love her dearly.  So, my situation model also includes an olfactory representation as well as a sense of affection for the dog. 

          From this brief description we can already illustrate the components of a situation models.  These components include representations of entities (dogs, fireplaces, wet dog smell, etc.), properties of entities (the blackness and wetness of Abby‘s fur, my affection for her, etc.).  These components are structured in accordance to spatial-temporal frameworks (the dog in front of the fireplace, etc.), and relational information (my ownership of the dog, my sense of her as part of my family etc.).  My situation model is different from that of the person with the German shepherd in their model, yet both situation models are appropriate.  Again, situation models are in an appropriate range when they depict plausible situations based upon the typical uses of the words that cue them.  Of course, this may admit of degrees. If my situation model had a wood stove instead of a fireplace, it would be less appropriate since ‘fireplace’ does not typically denote such a thing. An inappropriate situation model flouts typical usage of the terms that cue it.  A situation model that includes a frog at Niagara Falls is inappropriate.

          An experience is developed through this situation model.  I can see, smell and feel affection for the dog through my situation model.  Of course this is a “guided” experience^[23] and it is limited by the description given.  In our example the dog could be any dog, but that is not necessarily the case.  If the description had been of a Great Dane in front of the fireplace, only a situation model containing a Dane would be appropriate.  As a consequence, my experience would be different.  For me, the experience would lose the sense of attachment since I have no attachment to any Great Dane.

          Here is another example that is explored by Rolf Zwaan in “Embodied Cognition, Perceptual Symbols, and Situation Models.”^[24] Zwaan offers an explanation of how to ride a mountain bike over a log and then discusses what representations might be cued by the description.  Here is the description to be discussed:

(2) Make sure you are in a gear that is not too high and not too low.  Make sure you have sufficient speed and that you are perpendicular to the log.  Then get up out of the saddle, keep your feet still, and lift the front wheel so that it clears the log.  Your large gear will hit the log and your momentum will carry you across the log.  Stretch your arms before the front wheel hits the ground so that you don’t flip over.  Start pedaling as soon as your front wheel hits the ground.^[25]

A situation model for the above paragraph will include a representation of a rider or protagonist, and his/her position and orientation.  Also represented would be the bike, log and trail and their relationship to one another.  One would also represent the sequences of events.  The structure and components of one’s situation model will need to be continually updated in accordance with the new material presented.

          If we take into account the possible complexity of dispositional representations, we might also find that associated sounds and emotions are represented.  For the experienced mountain biker, the situation model would be richer and finer grained.  As Zwaan explains, “…the expert might be able to mentally simulate the position of his or her limbs. The expert might have some sensation of the speed at which the trees along the trail will pass (optic flow) and the speed at which the log approaches.  …The expert might “feel” how his or her hands are curved around the handles and the pressure of the pedals on his or her soles.”^[26]  If this is true, it seems that the expert may be said to experience more via their situation model and have the possibility to be affected more profoundly by the description than the novice.  Of course, a person with a vivid imagination may have a great deal of success recreating this scenario for their situation model, but it will be much more difficult for them to do so.  The point is that the experienced biker has a set of readily available dispositional representations that will automatically be part of their situation model. 

To further prove this point, we should note that it has also been indicated that people who have experienced something similar to the event described are more likely to put themselves in the protagonist’s position and to feel similarly to the protagonist.  Those without such experience may take the protagonist’s position or may view them from a “third-person” perspective.  For example, studies have been conducted that show smokers and ex-smokers respond as though they were the protagonists when listening to a scenario about people smoking after-dinner cigarettes.  They have cognitive and physiological responses that are similar to smoking urges in smokers.^[27] This can be explained by the availability of dispositional representations to those who have experienced an event as opposed to those who have not.  The more one has smoked, the less effort it takes to create a situation model for this situation because the sensations necessary to such a representation are familiar. 

          In example two, it seems that we take the position of the protagonist.  I do not believe that this is necessarily the case.  Consider the following experiment presented by William Horton and David Rapp in, “Representing Perceptual Availability During Narrative Comprehension.”^[28] Experiments were designed to test the claim that we organize our representations around the experiences of the protagonist because we adopt her/his perceptual experiences.  Thirty participants were tested.  They read experimental stories that contained information both blocked and unblocked from the protagonist’s perspective. This information can be “blocked” or “unblocked” by information that follows it in the narrative.  This information can serve to inhibit the protagonist’s awareness of prior information. Questions about the story were designed to determine whether or not unblocked information was more accessible to the reader than blocked information.  Target information is the information that the experimenter hopes that the reader will retain.  What follows is a sample story from an experiment:

(3)  Mr. Ranzini was sitting outside on his stoop.

He had lived on this block for 30 years.

Next door was a local playground for the children.

Directly across the street was the mailbox that he used. (target info.)

As usual, Mrs. Rosaldo was taking her poodle for a walk.

{Before test point}

Blocked story continuation

Suddenly, a large truck pulled up in front of

Mr. Ranzini.

The man in the truck was singing loudly to himself.

Unblocked story continuation

Suddenly, a man on a bike rode up in front

of Mr. Ranzini.

The man on the bike was singing loudly to

himself.

{Blocked/Unblocked test point}

Target Question [Experiments 1 & 2]:

Was there a mailbox across the street?

Control Question [Experiment 2]:

Was a playground for children next door?^[29]

          Researchers found that readers were slowest to verify the presence of target information when intervening story events had blocked it.  However, it is possible that the blocking events are particularly salient, reducing the accessibility of any prior story information, regardless of perceptual availability.  So, while it seems true that we are representing the events spatially as we read them, this does not mean that we are strictly adopting the protagonist’s perspective.  We may merely be modeling information from a third-person perspective.  The authors of this study accept that this is a possibility, while leaning toward the conclusion that we do put ourselves “in” the protagonist’s position.  Either result acts as support for the idea that we do represent spatial and visual information in our situation models and this is a positive result for the idea of situation models in general and how and what they represent.

          We do not have to put ourselves in the protagonist’s position to create a guided experience via a description or a narrative.  Our situation models will be influenced by our own dispositional representations and experiences that in turn will determine the cognitive effect of the narrative.  I may experience the protagonist’s emotions in a direct way or experience a different, sympathetic emotion as an observer.  This explains the varied responses that people can have in response to narrative.  While we all will represent certain aspects of a narrative, elements like mailboxes, children playing, bicycles, logs, trails, etc., we will represent them differently or idiosyncratically due to our own knowledge and experience with such elements.  Some people will have emotional experiences, gain insights, etc.; some experience less affect or none.

          A final example to consider:

(4)  Imagine going on a spelunking expedition. You lower yourself into a hole in the ground and enter a dank, winding passageway. After a couple of bends there is absolute pitch darkness. You light the carbide lamp on your helmet and continue. The passage narrows. You squeeze between the walls. After a while you have to stoop, and then crawl on your hands and knees. On and on, for hours, twisting and turning and descending. Your companion, following behind you, began the trip with enthusiasm and confidence; in fact she talked you into it. But you notice an increasingly nervous edge in her voice. Eventually, the ceiling gets too low even for crawling; you wriggle on your belly. Even so, there isn't room for the pack on your back. You slip it off, reach back, and tie it to your foot; then continue, dragging the pack behind you. The passage bends sharply to the left, as it descends further. You contort your body, adjusting the angles of your shoulders and pelvis, and squeeze around and down. Now your companion is really panicked. Your lamp flickers a few times, then goes out. Absolute pitch darkness. You fumble with the mechanism . . .^[30]

          What I find interesting about this example is that it illustrates the possible impact of our situation models on our current and future cognitive states.  A situation model representing this dark, dank scenario can call matters of our underlying character into play.  If one is claustrophobic, one will have a strong response to the idea of squeezing through the tunnel as the pathway gets smaller and smaller.  A person in such a situation may even start to consider options such as turning around and going back to the beginning of the cave even though that possibility is not mentioned in the narrative.  This is the kind of case that leads me to believe that situation models create an experience.  We start to feel and think as we would if in the actual situation.

          One might also start to feel claustrophobic imaging this situation even if one is not usually claustrophobic.  This could transfer into other situations in real life if the effect is strong enough.  One may begin to feel uncomfortable in elevators, etc., when that was not previously the case.  This possibility runs parallel to a discussion of the movie Jaws in Richard Gerrig’s Experiencing Narrative Worlds: On the Psychological Activities of Reading.^[31]  Gerrig reports several accounts of people who came to fear the water and sharks after viewing this film - one of them a diver.  I also know a person who no longer will swim in anything but a pool since she saw this film 30 some years ago!  This evidence is anecdotal, but it occurs often enough to warrant some consideration.

          One question that arises is whether we actually have to have a mental image to understand a narrative.  It is not my experience, and I expect there are others who would agree, that I always completely picture a situation while it is being described to me.  It is possible that we may be constructing situation models without completely attending to them.  This seems an odd claim, but if we can construct our dispositional representations without consciously trying to memorize the elements involved, it seems possible that we could activate these representations without being entirely conscious of their activation.  Also, there will be elements to any representation that may not apply in the context of the current description or situation model.  Maybe part of my representation of Abby is the sound of her bark, but it is not currently my focus.  Further, if there can be degrees of active awareness of the construction of a situation model, this could explain differing degrees of comprehension and effects of the narrative.  I do not get as many of the implications of a narrative if I am not actively attending to the construction of my situation model.  This would also explain varying degrees of the impact of narratives on different comprehenders.  If I am actively engaged in constructing a situation model, I am more likely to feel involved with and attached to characters.  For example, when the protagonist suffers tragedy, I may experience some sadness on her behalf if I am involved in the situation model.  If I am not attending to the experience, I may not feel much at all. 

          It also seems to follow from the theory of situation models presented here, that more familiar situations can be represented without conscious attention to their representation while less familiar situations will require attention to the model.  This may also explain why we are not consciously aware of all of our situation models.  The more familiar the situation, the more automatically a representation can be constructed without my awareness.  If I think of Abby often, I will not need to concentrate on recreating her as I would if I rarely thought of her.  Building on the analogy with learned skills, the more often I activate a firing pattern, the less conscious I am of its activation when it is cued.  This suggests that we need to more consciously attend to the creation of situation models for narratives that are unfamiliar in some way.  We must be more active in creating situation models for experiences that are not like our own or include elements with which we are unfamiliar.  For example, it would be much more difficult to create a situation model for driving a car from reading a description if I have never done it or seen someone else do it than if I have had the experience myself.  If this is true and it is also true that the more we attend to our model, the more impact it will have on us, then it may also be true that we would be most impacted by more novel or unfamiliar narratives.

          The fact that our representations are dispositional might help to explain why people can understand the same narratives differently.  Two different people may focus on different aspects of a narrative resulting in slightly different understandings of the narrative.  For example, many people read The Wizard of Oz and represent it as a fairytale about a girl who learns to appreciate the beauty of her home, family, and friends because they are more attentive to the relationships between characters.  Others may read it as a tale about the battle between good and evil because they focus on different behaviors exhibited by the characters.  Children, who have less experience with narratives and life in general, may simply represent the situations described and nothing more.

          From example three, we see a description/narrative that can possibly change one’s response to experiences that occur in our everyday lives.  Can situation models explain this possibility?  It is believed that in everyday experiences representations that occur together often become strongly associated.^[32]  When we build situation models we are activating a variety of representations simultaneously.  It seems plausible that the more representations occur together in the maintaining and updating of this situation model, the greater the chance that they may become associated.   It seems possible that a dispositional representation, through its activation in a situation model in response to narrative, can gain new associations with other dispositional representations in that situation model.  The more often two associations are activated together, the more likely they are to be associated with each other in the future.  This could potentially affect the way symbols are framed or organized in a schematic representation. 

Toward a theory of metaphor as guided experience

 As stated earlier, my main interest lies in claims that metaphors can give us insight, change how we think and feel about something, and even change our very cognitive structure.^[33]  How can the idea of language comprehension as guided experience lead us to an explanation of metaphors’ more befuddling effects?  The first step to answering this question is to acknowledge that metaphors rarely occur in isolation.  We encounter metaphors in the context of a conversation, narrative, or even in expository texts.  If we accept that we build situation models in response to linguistic cues, and that metaphor occurs in the context of these cases, it follows that metaphors are to be understood in the context of an active situation model. 

          Let us first consider unique metaphors, which tend to require some thought before we will claim to understand what they mean.  When we encounter a novel metaphor, we encounter a segment of language that cannot automatically be integrated into our situation model.  The typical usage of the words in a metaphor does not work to create a situation model that makes sense or mirrors the logic of the world as we experience it.  This being the case, we must actively and consciously consider novel metaphors in order to place them in the context of the situation model.  Integrating a metaphor into our situation model is a process that I think explains the answer to some long standing questions regarding how metaphors are interpreted. 

          What does integrating a metaphor into a situation model require?  To answer this question I will build on some ideas offered by Max Black in, "Metaphor."^[34] Black claims that a metaphor has two subjects: the principal and the subsidiary subject.  He describes these subjects as 'systems of things' or commonplaces associated with a subject rather than the things themselves. To interpret a metaphor we attribute the relevant commonplaces of the subsidiary subject to the principal subject.  Black's theory has been criticized for not sufficiently explaining how the relevant commonplaces are determined and also for failing to provide us with a reliable way of knowing which subject is the principal subject and which is the subsidiary subject.  When a metaphor occurs in ‘a is b’ form this determination is easy, but many metaphors resist translation into this form (as in the example, “Then how should I begin to spit out all the butt-ends of my days and ways?”)  In order to use Black’s basic ideas I must first change the idea of applying commonplaces from one subject to another to the application of associations, as described above, from one subject to another.  Once this change is made, we can use the situation model theory to explain what Black left unexplained.

          First, associations should be redefined as dispositional representations that are cued by the mention of a concept.  Black uses the idea of commonplaces to indicate associations common to a society or culture, rather like stereotypes.  This understanding of associations will still be the case to some extent, since we form our dispositional representations in the context of a common society or culture.  But there is more to associations on my account. Associations will be multimodal and can evolve and change over time.  Each person’s associations for a concept will have some general agreement due to similarity in our perceptual experiences, which result from our living in the same world, but each person’s representations will have idiosyncrasies.  All of those who know my dog Abby will have some overlapping elements to their representations, but some may represent a slightly smaller looking Abby than others or some may represent an Abby with more rambunctious behavior than others.   

          How are appropriate associations chosen for a situation model?  The perspective and current context of the situation model determine the appropriate associations.  First, the perspective we are directed to adopt will eliminate some of the associations that might tend to occur with a certain term or the concept it represents.  In the context of The Wizard of Oz, for example, when Dorothy first encounters the Wicked Witch of the West we tend not to consider the grief that the witch may be experiencing at the loss of her sister.  This seems unusual since often when we hear a story about someone who believes her sister has been murdered, this emotional state would be a relevant association.  Instead, because we have taken the perspective of Dorothy in constructing our situation model, we focus on the fear that Dorothy must feel at gaining such an evil and powerful enemy.  In fact, it almost seems ridiculous to consider the Witch’s feelings even briefly. 

          Second, the context helps determine appropriate associations.  If a metaphor arises in the context of a conversation or text, perspective is already developed before the metaphor occurs.  We are already inclined to consider certain associations and not others because the context has chosen the most likely activations.  Certain dispositional representations are more likely to be activated than others.  When in the context of Romeo and Juliet we encounter “Juliet is the sun,” we have already constructed a model that puts us, in this instance, “in” Romeo’s perspective, or at least a mediated perspective similar to Romeo’s, and that has chosen our attitude about Juliet to be a romantic one.  Romeo regards Juliet as the principal subject.  Like Romeo, we are already focused and probably have already activated representations that highlight her best qualities.  In this context, the metaphor activates associations of warmth, beauty, and light that are likely to emerge.  In an appropriate situation model these are within the appropriate range of associations.  If an association is within the appropriate range, one should be able to go back to the original situation model and explain why these associations make sense while others do not.  An association like, “ball of hot gases,” is an unlikely association to be activated due to the context.  Such associations have no basis in an appropriate situation model for this work.  Were this to arise in a different context, one where we were focused on Juliet’s less attractive qualities as a know-it-all loud mouth, these associations may become relevant.  In the context of Romeo and Juliet, this association is not as available in our currently constructed situation model.

          It may be argued that these associations could be cued easily by literal language, so metaphor adds nothing special.  What I think gives metaphor some power to impact us beyond the literal statement is that we work to integrate it into the situation model at a more conscious level - or at least we do if we want to claim some understanding of it.  Integration is less automatic with figurative language.

          Often we must stop and work through a metaphor before it makes sense to us.  Consider the following example of the metaphorical statement, “I wonder if General Motors’ health is improving?”^[35]  We may work through such an example by trying associations from both terms together to make them fit.  In the context of a situation model built for a discussion of the current economy and growth possibilities of different companies, associations for “health” may include different illnesses and indications of improvement.  We may work through this metaphor by fitting together its associations. For example, we may think of illness as low profit and little money for the development of new products to create growth.  Improvement would mean a positive change in these associations: more profit and greater growth potential.  I think when we are really attempting to comprehend a novel metaphor and integrate it into a model, we do work this way.  I have spent many afternoons listening to the English Professor I shared an office with do just this with students from his poetry class.  Students were encouraged to think through associations and which ones fit into the context of the poem.  Through this process students would find and/or create interpretations. 

          Many theories of metaphor suffer from the problem of positing some automatic comparison or transfer between referents, meanings, or schemas of a metaphor’s terms.  This seems to leave us with two important problems.  If this process is automatic and unconscious and is based solely upon the meanings or contents of terms that we all have access to, everyone who encounters the metaphor ought to be able to interpret the metaphor easily and be said to understand it.  As indicated earlier, I suspect that the process is not automatic or immediate.  Whether we do it consciously or not, we must do some cognitive work in order to interpret a metaphor.  It is true that we encounter the occasional metaphor that others seem to comprehend that we do not or vice versa, even when we all should have the appropriate representations with which to work.  The missing element may be the desire or ability to do the work.  It also occurs that the same metaphor can be interpreted slightly differently from person to person.  This can only be explained by a theory of comprehension that allows for the definition of terms to be idiosyncratic or the processes of interpretation to be idiosyncratic as a theory of guided experience could.  The above process is an example.

          Once we can fit the relevant associations into our situation model, we can then be affected by the metaphor.  Some of the effects that are attributed to metaphor are: the sense that a new insight has been gained, that something is seen in a new way, and sometimes emotional responses.  How can this be explained on this view?  It could be that we are creating new associations between previously unrelated dispositional representations.  If we work through a metaphor in a way that might cause the simultaneous activation of dispositions that have never occurred simultaneously before, this is a genuinely new cognitive experience.  Concepts that were unrelated are now, if only momentarily, related.   Further, it seems possible that some dispositional representations carry associated emotional responses that may now occur with a previously unemotional representation.  This new association may cause us temporarily to feel emotion toward things we usually would not.  Further, experiencing, in this guided and somewhat limited way, the situation may also explain emotional responses to the situation modeled. 

          To give an overview of this section, on this theory, a metaphor is characterized by the fact that it presents us with material that does not immediately fit into our working situation model.  We can find associations that are relevant in our situation model with some effort.  The perspective taken in the situation model and the context of the situation model in which the metaphor occurs will determine appropriate associations.  The best, most appropriate, associations are the ones that make the most sense in the context of the working situation model.  The effort is worthwhile because we will tend to consciously focus on the metaphor and finding its meaning.  This extra effort often results in extra cognitive impact since we will process the information in the way we do when we are learning something new.  As with the driving example discussed earlier, when we are learning a new skill or new information, we must pay attention.  After a while, the skill becomes automatic and its activation often goes unnoticed.  The initial effort results in a cognitive effect.  With metaphor, the effect is that we may come to permanently associate two things that we previously did not and would not merely through everyday experience.  As with the ‘argument is war’ category of metaphors, we may even come to behave in accordance with the metaphor without even noticing that we are doing it.

Exploring a complicated example

In order to bring the above ideas together, it may be helpful to discuss an example.  Below is the first stanza of Wordsworth’s “I Wandered Lonely as a Cloud”.  Poetry is a particularly difficult case due to the number and novelty of the metaphors involved.  The strength of this theory of metaphor as guided experience is that it gives us a perspective from which to interpret and whittle away relevant associations.  This exploration will also illustrate the requirement that novel metaphors require the intent to interpret and cannot be explained by automatic response to literal meanings.

I wandered lonely as a cloud
That floats on high o'er vales and hills,
When all at once I saw a crowd,
A host, of golden daffodils;
Beside the lake, beneath the trees,
Fluttering and dancing in the breeze.

          This passage contains several metaphors as well as a difficult to interpret simile.  We first encounter the simile while we have no situation model in place to aid in its interpretation.  We must reserve interpretation for after we have a situation model and this model will depend upon the interpretation of other metaphors.  This may require another type of representation of the simile until it can be integrated into the situation model.  Another possibility, the one I think the more likely, is that we will represent all associations and some will fall away as they are shown to be irrelevant.

          We do have noun phrases to cue dispositional representations, which are the basis of our situation model.  We will represent the cloud, a golden daffodil, a lake, trees, the breeze and the relationships between these things all from the perspective, or mediated perspective, of the narrator as a cloud.  All of these representations will be idiosyncratic and will be of differing complexity due to these idiosyncrasies.  As with Zwaan’s mountain biking example, those who have experienced such a scene directly will have more detailed situation models and the model will be constructed more easily for them. 

          We must interpret how to represent the ‘crowd’ of daffodils since this is not the usual way of describing flowers but is commonly associated with people.  Crowds have the associations of being large in number, sharing one attitude, being either organized or unruly, etc..  The description of a field, seemingly untouched by humans, suggests that the appropriate associations of ‘crowd’ are the large number, unmanaged and unorganized.  But crowds can have a variety of attitudes and which fits here still needs to be determined.  Once we encounter the description of the daffodils as “fluttering” and “dancing” we can further interpret the character of the crowd.  Dancing has a variety of associations as well.  Dance can be expressive of an emotion or it can be inexpressive.  The varieties of emotions possibly expressed are numerous. The idea of “dancing” could be represented as active and aggressive or peaceful.  Flowers are not usually associated with aggression and breezes are gentle as opposed to gusts of wind, so we can determine that the dancing should be understood as a joyful movement.  Other representations will be disregarded.  The crowd now seems to be a happy, pleasant gathering, and associations of rowdy angry crowds will no longer be part of the representation. 

          Now we can interpret what makes the cloud lonely and how the narrator may be like this lonely cloud.  Clouds have associations such as fluffy, white, above the ground, and in constant motion through the sky. In this case, the distance of the cloud from the scene on the ground seems relevant as does the cloud’s inability to join in the scene.  Fluffy and white do not seem relevant.  The cloud must watch the pleasant gathering from a distance.  It is an outsider, as is the narrator.  

          The effect of such a narrative and the metaphors included therein will differ from person to person.  One who usually considers such natural scenes as uninteresting or lonely may construct more positive associations with such scenes because they can now see some aspects of nature as a celebration.  They may even react to them differently in the future.  Those who already regard such scenes in this way will probably experience less affect.  But some people may come to notice the source of loneliness as impermanence and detachment from others and this may change their dispositional representations of loneliness in the future. 

Conclusion

How can this theory of metaphor as guided experience help to explain what other theories have left unexplained?  One of the main strengths of this theory is that it accepts that the literal meaning or typical usage of terms is necessary for understanding a metaphor.  It is not the meaning of terms that is important to us though.  What is important is the juxtaposition of these uses or meanings and how they relate within the context of a situation model.   This helps us to explain what the effect of a metaphor is.  Metaphors can create associations between terms not usually associated.  Most often these associations only last as long as the situation model exists.  If representations are associated often enough, as with dead metaphors, then that can affect a change in the dispositional representations themselves.  This can explain some of the common metaphors and the resulting behaviors that Lakoff and Johnson use to show the influence of metaphors on our concepts themselves. This effect is not expressible in propositional form because it is often a restructuring of the cognitive structure, which cannot fully be explained.  The effects may be endless and even unknown to the affected.

          Comprehending a metaphor is a matter of integrating it into a situation model, which allows us to determine the appropriate associations to be used in the understanding of a metaphor.  The perspective we assume in constructing our situation model chooses the principal subject of a metaphor and then we actively apply associations from the subsidiary subject to the principal.  We can see how context, in the form of a situation model, provides perspective that enables us to narrow down the options of possible interpretations for a metaphor.  It is also plausible that a particular situation model may cause us to create similarities, where formerly there were none, by giving us a new perspective on an old representation.  It also does not overly restrict the interpretations.  Because our dispositional representations are idiosyncratic, so is our interpretation of metaphors (and literal language for that matter).  This allows for a range of appropriate interpretations.  We can also see that interpreting metaphors requires some active participation in the process of integrating the metaphor into a situation model.  If we do not make some conscious effort toward integration, we will not understand the metaphor. 

Works Cited

Barsalou, Lawrence J.  “Perceptual Symbol Systems.” Behavioral and Brain Sciences 22 (1999): 577-609.

Black, Max.  “Metaphor.”  Models and Metaphors.  Ithaca: Cornell University Press, 1962.

Black, Max.  “More About Metaphor,”  Dialectica. 31 (1977).

Damasio, Antonio.  Descates’ Error.  New York: Grosset / Putnam, 1994.

Davidson, Donald.  “What Metaphors Mean.”  Inquiries into Truth and Interpretation.  Oxford: Claredon Press,1974.

Drobbes, D.J. and Tiffany, S.T.  “Induction of smoking urge through imaginal and in vivo procedures: Physiological and self-report manifestations.”  Journal of Abnormal Psychology 106 (1997): 15-25.

Gerrig, Richard.  Experiencing Narrative Worlds: On the Psychological Activities of Reading. New Haven, CT: Yale University Press, 1993.

Kittay, Eva Feder.  Metaphor: Its Cognitive Force and Linguistic Structure. Oxford: Clarendon Press,1987.

Kittay , Eva Feder. "Metaphors as Rearranging the Furniture of the Mind: A Reply to Donald Davidson."  in A Metaphorical Point of View.  ed. Z. Radman, New York : W. de Gruyter, 1995.

Lakoff, George and Mark Johnson.  Metaphors We Live By.  Chicago: Chicago University Press, 1980.

Lakoff, George and Mark Turner.  More than Cool Reason: A Field Guide to Poetic Metaphor.  Chicago: Chicago University Press, 1989.

Walton, Kendall.  "Spelunking, Simulation and Slime: On Being Moved by Fiction". In Emotion and the Arts. eds. Mette Hjort and Sue Laver,  Oxford University Press, 1997. 

Zwaan, Ralph.  “Embodied Cognition, Perceptual Symbols, and Situation Models.”           Discourse Processes 28 (1999): 81-88.