Sunday, 31 May 2009

Mental Model Literature Review - Part 2

In the previous post I began with attempting to shape my understanding of Mental Models by looking at some of the seminal books and papers often cited. I have not yet personally arrived at any full understanding of mental models nor how to use them to explore my primary research into Design Models. In this post I want to explore some of the language and concepts that appear to be common to the study. One of the reason's for doing this is not just to understand the language from another discipline, but to be able to use the language to fix my research position in cognitive science, before discussing Indy Young's interpretation of mental models within interaction design (Young, 2008).

Again, for the casual reader of this blog, I must state that this post is simply an exercise in synthesising my research to help me understand this complex subject. This post is not a finished thesis by any means, and shouldn't be read as such. I do aim to formalise this research into a paper to underpin my first practical PhD project later on. So feel free to comment, suggest other interpretations, disagree etc.


‘Why do people use mental models?’ We think mental models assist human reasoning in a variety of ways. They can be used as inference engines to predict the behavior of physical systems. They can also be used to produce explanations or justifications. In addition, they can serve as mnemonic devices to facilitate remembering.”

(Williams et al, 1983, p135)

"Mental Models are simply affinity diagrams of behaviors made from ethnographic data gathered from audience representatives. (...) A mental model for a particular topic is, in essence, an affinity diagram of user behaviors."

(Young, 2008, pp2-3)

Spanning 25 years the language as to what constitutes a mental model has naturally expanded, contracted, morphed and become specialised. But is the term still describing the same cognitive concept? There appears to be be a marked difference in the use of the term at present within interaction design than the term used by Cooper, Norman and those within the cognitive science discipline. Building upon the research documented seminally in 1983, mental models at present seems to have two appearances: Indy Young's 'affinity diagram of user behaviours' (2008) and 'cognitive modelling to aid understanding'. Alan Cooper, author and chairman of interaction design company Cooper, sees things more from a more user-centred perspective akin to the cognitive scientists definition. Cooper, Norman and others understand mental models as a personal tool used by each individual to make sense of a situation, task or encounter. Cooper says,

“A person’s mental model is their own internal representation of reality – the way they think about or explain something to themselves. Mental models are deeply ingrained and are often the result of a lifetime of experience. People’s expectations about a product and the way it works are highly informed by their mental model.” (Cooper et al, 2007, p118)

Therefore to enable me to reach a clear definition of mental models, that will lead me onto my research into the Design Model, I will now explore the language found within the literature in order to resolve the Young/cognitive science interpretation of the term. I will explore Young's idea and methodology in another post.


One of the aspects of mental models is that humans recognize structural patterns based upon previous experience or understanding of previous functionality. Mental models cannot and do not emerge fully formed, they are dynamically conceptualized from components of previous experience and knowledge. The individual selects the closest fitting components that together give them a model to understand the causality of the interaction. These components of experience/knowledge come from a prior learning process, and will themselves evolve into better components due to familiarity with the interaction over time. These components that are dynamically placed into a causal linear context are said to be cognitively "runnable". The person can conceptually "run" the causal components in their minds to envision the actual interaction, its process and its consequences. The term interaction referred to here is a general term. Interaction is a term that has an identity issue due to the different disciplines involved in defining it. Dr. Sally McMillan identified this problem nine years ago,

"while some scholars see interactivity as a function of the medium itself, others argue that interactivity resides in the perceptions of those who participate in the communication."
(McMillan, 2000)

It could be argued that the very nature of the range of definitions for interaction enriches the research surrounding its possibilities, but it can also obscure unless the exact parameters are defined to contextualise what form interactivity takes. One current industrial interpretation by Hugh Dubberly et al, of Dubberly Design Office (DDB) in San Francisco, of interaction approaches it from a systems-theory point of view (Dubberly et al, 2009). In an article posted on the DDB website he and his colleagues explore a broader systems interpretation of interaction. A paper I have previously found useful is one by Dr. Spiro Kiousis which was an explication of the term. In "" Kiousis suggests that

"interactivity is both a media and psychological factor that varies across communication technologies, communication contexts, and people's perceptions."
(Kiousis, 2002, p355)

In his explication he identifies a literature review, which by current standards a good historical benchmark, of types of interactivity. Through this he concludes that there are three common factors to interactivity from an operational point of view:

"technological structure of the media used (e.g. speed, range, timing flexibility, and sensory complexity), characteristics of communication settings (e.g. third-order dependency and social presence), and individuals’ perceptions (e.g. proximity, perceived speed, sensory activation, and telepresence)."
(Kiousis, 2002, p379)

Dubberly et al in the intervening years explore the transfer function of interaction. Looking to both Norman and Verplank Dubberly et al explore the dynamism of interactivity,

"in ‘interaction’ the precise way that ‘input affects output’ can itself change; moreover in some categories of ‘interaction’ that which is classed as ‘input’ or ‘output’ can also change, even for a continuous system"
(Dubberly et al, 2009, p3)

They refine this dynamic definition of interactivity to cover six systems: reacting, regulating, learning, balancing, managing & entertaining, and conversing (see diagram). Now I don't want to stray too far in discussing the definitions of interactivity in this post too much, I feel it is important if I am to make sense of a user's mental model being constructed to understand the interaction they are involved in. The component models that a user will "run" will naturally follow their previous experience, which in turn will informed by a form of interactivity they previously encountered. Therefore to understand a mental model dynamically conceived is contingent on also the form the interaction takes. Here is where Norman's term 'system image' twenty five years ago was coined to describe the cognitive process of how a user identifies the causality of the interaction. Each component model a person joins together to make sense of the interaction can be viewed as an 'object' with its own autonomy, its own properties are invariant and modular so that when dynamically formed as a mental model to understand one problem, that model can be disintegrated dynamically and rebuilt into a new model for a different problem using the same cognitive components. These component models of understanding are often naive, fallible, incomplete and inaccurate under quantifiable or qualitative scrutiny. They are based upon perception, learnt behaviour, deduction and previous experience of affordances. Therefore they are empirical and heuristical (Williams et al, 1983).

To understand a phenomenon the person establishes a proposition as to what needs to be achieved. This proposition is arrived at through the affordances that the phenomenon suggests. Leaving affordances to one side while I continue to nail down the roots of understanding the formation of a Mental Model. I will still be quoting from work done twenty five years ago, still cited today as seminal works, in order to then cross-examine with newer ideas as to what constitutes a mental model. Johnson-Laird was the first describe the cognitive process from a propositional reasoning point of view. Beginning with a hypothesis that people form visual images of objects and scenes by using their imaginations and experiences to dress the images, Johnson-Laird identifies that there are two psychology schools of thought upon the understanding of images: the 'imagists' and the 'propositionists'. The imagist school argue that images are a distinct form of mental representation, while the propositionist school argue that images are a secondary phenomenon, that images are a 'single underlying form of mental representation, and that images merely allows the stored experiences from previous phenomena to be more easily accessed. People who draw on this store are not adding any new information in the construction of an image. It is secondary to and underlying the main propositional representation from which an image is created. It is crucial to state that propositions in regard to mental representation are boolean, abstract and their structure are not "analogous to the structure of the objects they represent". Mental representation can therefore be broken down into three forms. The propositional representations are strings of symbols strung together to form the syntax that elucidates the mental model. The mental model is a structural analogue of the phenomenon, while the image are perceptual correlates, complimentary visualisations created from mental representations of the model from a particular point of view (Johnson-Laird, 1983). Johnson-Laird describes the distinctions between them as high-level mental processing, and ultimately the construction and manipulation of a mental model is to make it possible for a person to "reason without logic". To explain this I wish to quote Johnson-Laird's explication on propositional reasoning before I end this post.

“Philosophers have generally taken propositions to be conscious objects of thought – those entities that we entertain, believe, think, doubt, etc., and that are expressed by sentences (…). Since I am concerned, not with the nature of the ‘machine code’ of the brain (…), but with the types of higher level of representation, I propose to revert henceforth to the traditional philosophical terminology: a propositional representation is a mental representation of a verbally expressible proposition. (…) Propositions can refer to the world. Human beings, of course, do not apprehend the world directly; they possess only an internal representation of it, because perception is the construction of a model of the world. They are unable to compare this perceptual representation directly with the world – it is their world (…). Propositions can also refer to the imaginary or hypothetical worlds. One proposition may be false of such a world given that others are true of it. Human beings can evidently construct mental models by acts of imagination and can relate propositions to such models. (…) Unlike a propositional representation, a mental model does not have an arbitrarily chosen syntactic structure, but one that plays a direct representational role since it is analogous to the structure of the corresponding state of affairs in the world – as we perceive or conceive it. However, the analogical structure of mental models can vary considerably. (…) A characteristic difference in the contents of mental models, images, and propositional representations, concerns their specificity. Models, like images, are highly specific – a characteristic which has often drawn comment from philosophers. (…) Hence, if you reason on the basis of a model or image, you must take pains to ensure that your conclusion goes beyond the specific instance you considered" (1983, p155-158).

I will end this long post here as I wish to take another approach to understanding a mental model in my next posts. Again let me leave you, the reader, with my purpose for this post. It isn't a completed thesis by any means, merely a representation of the process I am taking to synthesize and understand a very complex subject. My interpretations, assumptions and understanding so far may be wholly wrong, naive or even accurate. This post in its current state should merely be read with my intentions in mind. Any comments, suggestions etc. would greatly be appreciated, especially if you are from a cognitive science discipline.



DUBBERLY, H., HAQUE, U., and PANGARO, P. (2009) What is interaction? Are there different types? Dubberly Design Office [online], [Accessed 15 March 2009], (pp.1-12) Available from World Wide Web:

KIOUSIS, S. (2002) Interactivity: a concept explication. New Media & Society, SAGE Publications, Vol4(3):355–383

McMILLAN, S.J. (2000) ‘Interactivity is in the Eye of the Beholder. Function, Perception, Involvement, and Attitude Toward Web Sites’, in M.A. Shaver (ed.) Proceedings of the 2000 Conference of the American Academy of Advertising, pp. 71–8. East Lansing, MI: Michigan State University.

See Design Model - statement for a full list of references this post cites beyond those indicated above.

No comments: