In theoretical cases, too, the content of a model appears to outrun mere structure. The mass-flow equation, for instance, seems to express facts about pipe pressure—not merely variables. And indeed, when models do say things about the world, they do not appear to do this in a straight-forward way. There is no simple mapping between abstract properties and concrete properties  see Frigg ; Levy ; and Thomson-Jones for more detailed treatments.
By taking models as concrete, model-to-world relations can be understood as comparisons between or descriptions of concrete properties. Moreover, the non-mathematical features of models are easily accommodated. To fill out their views, fictionalists appeal to positions developed in aesthetics.
Models as Make-Believe: Imagination, Fiction, and Scientific Representation
In effect, taking models as fictions provides a conduit for applying conceptual machinery about fictions to models. This allows philosophers to attempt the tightrope between staying true to modeling practice while avoiding costly metaphysical commitments. A way into the philosophy of fiction is via notions of fictional truth. How can it be true when the actual world has never contained a little prince living on a very small planet?
The answer is to take the above sentence to refer to some fictional, rather than the actual, world.
1. Problems Concerning Scientific Representation
The challenge is working out where these internal truths come from. This has nice semantic features, but is rather problematic metaphysically speaking see, e. What do we take these possibilia to be? Children playing lava use material objects—furniture and the floor—as props. The rules of the game and the props prescribe such an act of imagining. Broadly speaking, in order to represent, an object must have propositional content. The semantic question asks how models represent in the context of fiction, the model-world question asks how models might be about the world.
The mechanics of truths about models can be straightforwardly transferred. This answers the semantic question. Similarly, The Little Prince is a partly allegorical work—from reading it, we are supposed to learn about the nature of, say, adulthood in the actual world. The world of adults in The Little Prince is absurd, and this absurdity is reflected in our own world.
We can categorize fictionalists by how they characterize this mode of theorizing. First, indirect accounts take the model-world relationship to be mediated by some further object—the model. Second, by direct accounts the model is, in some sense, a description: there is no mediation. To see the difference between direct and indirect views, a tripartite distinction first drawn by Ron Giere will be helpful see also Godfrey-Smith On this picture, we should distinguish between model descriptions , model systems , and target systems.
The model system is whatever it is that modellers directly examine. Answering the metaphysical question involves an account of model systems. A target system is the part of the world we take the model to be informative about. My father might specify a model system using a mass-flow equation, and take this to inform him about his target system, a pump. We might ask whether shifting the metaphysical burden from abstract objects to uninstantiated properties is less metaphysically onerous Levy More problematically, it is not obvious how comparisons between instantiated and uninstantiated properties can yield truths about the world Godfrey-Smith ; Salis Direct fictionalists aim for a cleaner metaphysical plate.
However, they differ on the metaphysics and model-world question. Where indirect fictionalists take a model system to be a game of make-believe, direct fictionalists, in a sense, deny there is a model system at all. Direct fictionalists suggest we take model systems to be descriptions of target systems.
My suggestion is that we treat models as games of prop orientated make-believe—where the props, as it were, are the real-world target phenomena. To put the idea more plainly: models are special descriptions, which portray a target as simpler or just different than it actually is. Levy This suggestion overturns the idea that models are indirect representations of target systems.
The immediate advantage is the avoidance of comparing either nonexistent fictional entities or nonexistent fictional properties with the existing world. There is simply the world and our interaction with it. This answers the metaphysical question, and internal truths are generated as they are for indirect fictionalists—via the props and rules of generation.
Arnon Levy explicitly targets the model-world question. The model just is a description of the world; the props are identified with the target system. However, the model is still a false description of the world: and so, how should we understand the model-world relation? Levy takes the relation to be one of partial truth.
Roughly, the notion of partial truth involves parceling up propositions or sentences into components, which are themselves true or false see Yablo So, the content of complex sentences can be understood as consisting in several parts, each with their own truth conditions. The last sentence is, of course, false—even in the fiction: Juliet is a 13 year old human female, not a star. A map gets it right when it gets it right in the relevant respects.
Models are direct, rather than indirect, representations. The purpose of going into this amount of detail about fictionalist views is two-fold. First, it is important to see that the two objections I provide in Section 5 are not solved by the kinds of precisifications attempted by Levy, Frigg, and company. On my view, both commitments are problematic in engineering or design contexts. Second, my aim in Section 6 is to encompass the successful parts of fictionalism into a broader account, and seeing how this operates requires a firm grounding in varieties of fictionalism. Rather, I deny that this aboutness is captured in fictionalist terms.
Drag to reposition
First, both direct and indirect fictionalists answer the model-world question in terms of there being some resemblance between a model and a fixed target system. For Frigg, this is a property-comparison; for Levy, a relation of partial truth. Second, for fictionalists, being a Walton-style representation is essential to modeling. But models are not representational in the same way that fictions are.
That is, they are not essentially representational. Stripped of representational content, a literary fiction is no fiction. That is to say, it is plausible that possessing fictional representational content is a necessary condition for something to count as a fiction. However, models may be—in fact often are — stripped of such content. Fictions qua fictions—as opposed to, say, inscriptions, or mistaken non-fictional reports, are always Walton-style representations. Models qua models are not.
During the pump selection process my father will often utilize models whose features are geared towards the specific site, and the particular pump he intends to select, so they are not general or targetless models. Determining pump suction requires taking atmospheric pressure into account, and as atmospheric pressure changes both with altitude, the position of the pump, and the depth of the well, these factors make specific differences to how we model the pump. However, the pump has not been selected as yet: the specific target appears to lack an extension.
What, then, does the model describe or represent? Presumably either a future object the pumping station that will be built or some kind of nonexistent object the pumping station intended to be built. What do these cases look like on fictionalist views? On an indirect view we are, presumably, comparing abstract properties to either future properties or further abstract properties. But these future properties are either indeterminate depending on how the world turns out , or our epistemic access to them is mysterious. On direct views, it is unclear what we should take the props to be. They could be, for instance, the well site—the object concurrent with the selection process.
Alternatively, they could be the future object—the completed pump. Perhaps we should view the model as a prediction about the future state; the model describes a future object.
- Models As Make Believe Imagination Fiction And Scientific Representation.
- Chinas New Deal: Economic Development Under International Financial Crisis.
- Annuaire Européen / European Yearbook: Vol. XIX;
- The War in Afghanistan.
However, we cannot take these models to be in the business of predicting future states. This is because they are preliminary : they are a way of getting the selection process off the ground. These early designs and models are often abandoned, functioning as scaffolding for yet more detailed models.
Such preliminary models typically succeed in their task: forming a basis for further design, necessary for the final product. The journey from an initial proposal to a complete pumping station is a long one, each step frequently involving the mediation of various proxies—and their abandonment. So, fictionalist construals of modeling answer the model-world question by positing a kind of relationship between the fictional content of the model and the nature of the target system.
- Clinician’s Manual on Intra-abdominal Infections;
- Power Versus Prudence : Why Nations Forgo Nuclear Weapons.
- Awakening to Awe: Personal Stories of Profound Transformation?
- Protected Blog › Log in;
- Solar Dynamics and its Effects on the Heliosphere and Earth: 22 (Space Sciences Series of ISSI).
- 2. General Griceanism and Stipulative Fiat.
Rather, the function of the model is to scaffold the construction of models which themselves are used for pump construction or selection. Moreover, there is often a dynamic relationship between models and targets in design contexts. A further objection to the idea that fictionalist accounts can provide a general, complete account of modeling practice arises from considering the models used in the last stage of pump-selection.
That is, for models to be models they must demand certain acts of imagining. Recall that the mass-flow equation, when taken as a veridical description of the world is, strictly speaking, false. First, the equation represents pipe friction as if it were not affected by, say, atmospheric pressure—it is non-veridical in a way similar to how the floor in a game of lava is not hot. Second, the output of the equation will be inexact: it will not predict pipe friction in a fine-grained way.
In different contexts we care more or less about each desiderata. After all, in such circumstances we only care that the model provides the appropriate output. In the third stage of pump selection, a pump is picked from amongst the standard options by, effectively, matching the curve generated by the pipe equation to the curve relating to the pressure output of each particular pump. This process is carried out automatically: a piece of computer software matches pump to well. This ensures that a pump is selected which is able to overcome the pressure, flow and friction from the required volumes of water.
But they do not do so as fictionalists require. That is, their functioning does not turn on their being props which demand certain imaginings in games of make-believe. It is hard to see how this could be so, given that the matching process is carried out by a computer. All they would need to do is match the curves as best they could. Such models could be used as props, but this makes no in-principle difference to their capacity to perform the task. That is, designing a model requires that we partake in the relevant game of make-believe via more or less explicitly specifying a set of rules of generation and constructing some prop.
It might be a necessary condition that, were the model to be taken as a representation in those terms, it would generate the relevant truths. But this is a weak condition: presumably any object could operate as any number of props in any number of games. Further, one might argue that non-fictional models are not models. This strikes me as unmotivated—my father utilizes a range of proxies to aid his selecting of a real-world pump. These proxies play a variety of roles, some fictional and others not, but all are surely part of the activity of modeling.
I imagine that many might be attracted to this last point, since such cases are not viewed as problem cases, but are simply excluded from accounts of modelling.
However, such exclusions seem ad hoc. If modeling accounts are supposed to be exhaustive, then we should prefer an account which happily accommodates them. We can take the case study which opened the paper, then, as involving three kinds of models, delineated in terms of their function. First, preliminary models. In pump selection, models are constructed in stages, preliminary models acting as scaffolds for more detailed, more targeted models.
Preliminary models are often used to aid in the imagining of the final product, but are not properly understood in fictionalist terms because their success does not turn on their saying approximately true things, or being approximately accurate representations, or possessing any other sufficiently detailed resemblance relation with a target system. Their role is to provide a basis to build further models in a dynamic interaction with the target system, and thus they play a crucial role in how the designed object turns out.
This is world-directed—the purpose of the model is to build something, after all—but its success in that world-directed function does not turn on its partial truth, or on property comparisons. Preliminary models make trouble for any account which cashes out model-world relations in terms of resemblance thus including many structuralist accounts. Second, fictional models. These, like the model which my father uses to determine the required properties of the future pump, do function as the fictionalist would have it.
Which is to say, the models act as props which urge games of make-believe by which the model says true things of the future pump. Third, procedural models —we see these in the final stage, where the optimal pump design is selected. Here, we simply care about its output: by matching the well system curves to pump pressure curves, the right selection is identified. We simply need to know that following this procedure will provide the result we need. The lesson is that—particularly in engineering—there are world-directed uses of models which fictionalists cannot account for. In this section I co-opt work from the philosophy of artifacts to argue that we should understand models as tools.
The content of a story is determined by the rules of generation and the props. The author controls these via how they design the props say, the words they write. Sometimes, my father might use the mass-flow equation in an explanatory context, for instance, in accounting for how some pump has malfunctioned. Here, a fictionalist treatment is attractive: the success of the model that it facilitates a good explanation turns on it having the right resemblance relation to the pump in question.
In other contexts, the mass-flow equation is a preliminary model. It seems reasonable to say that, when it is put to different uses, the content of the mass-flow equation changes. Content being relative to use is, as we shall see, a central component of artifacts. It is my claim that understanding models qua tools is deeper, more unified and more metaphysically kosher than understanding models qua fictions.
I will then attempt to answer the semantic and model-world questions by equating models with tools. Some philosophers particularly Amie Thomasson have argued that some apparently problematic objects such as fictional characters are abstract artifacts. On such views, fictional characters are something like platonic objects which bear concrete properties, but differ insofar as they are created.
Tools are a kind of artifact, specifically material objects which are used to manipulate other material objects. Tools are often designed, but are not necessarily so. Both a hammer and a rock can be used as a hammer. The suitability of a tool depends upon its material properties and the task at hand. Two kinds of content attach to tools, then.
But there are also what I will call success conditions. Some features of a tool are relevant to meeting success conditions and others are not. A set of tools frequently discussed by philosophers of science are instruments such as telescopes, spectrometers, and so forth. Although paradigm artifacts—sewing needles, hammers—are material objects, they are also intentional.
They are creations, ontologically dependent on their creators Hilpinen ; Thomasson By using a stone to hammer in tent pegs, I use the stone as a hammer. We can evaluate the success of an artifact on the basis of three criteria paraphrased from Hillipinen That is, did the creator make what they intended to? That is, whether the object the creator intended to make would be suitable for their purpose. Roughly, does the actual object do what we want it to? We can, then, distinguish between two classes of claims about tools: claims about material properties, and F- claims—claims which link the properties of an artifact to its purpose via the three criteria.
From Models-as-Fictions to Models-as-Tools
So, tools are a kind of artifact: they are built, intentional that is, functional objects. Tools have two kinds of properties: material features and functional features, F- properties. Material features are simply the properties of the tool qua material object.
Tools have a weight, a constitution, they behave in certain ways, and so forth. The vehicle is the medium through which the content is expressed. Its 2-dimensionalism is often an F -property, as the added simplicity allows better understanding of the relevant features of pump design. This move allows us to answer the semantic question and the model-world question, as follows. The semantic question asks us to understand claims about models.
1. Problems Concerning Scientific Representation
Rolf Hvidtfeldt. Home Contact us Help Free delivery worldwide. Free delivery worldwide. Bestselling Series. Harry Potter. Popular Features. New Releases. Description Scientists often try to understand the world by building simplified and idealised models of it. Adam Toon develops a new approach to scientific models by comparing them to the dolls and toy trucks of children's imaginative games, and offers a unified framework to solve difficult metaphysical problems and help to make sense of scientific practice. Product details Format Hardback pages Dimensions x x Illustrations note XVI, p.
Other books in this series.
- Photography and Science (Reaktion Books - Exposures)?
- VEP: J.P. Morgan Chase 2003 (Vault Employer Profile).
- Living Safely, Aging Well: A Guide to Preventing Injuries at Home.
- Account Options.
- Aquaculture and the Environment!
- Imagination, Fiction and Scientific Representation!
Philosophy of Epidemiology Alex Broadbent. Add to basket. Science in Metaphysics Vassilis Livanios. Models as Make-Believe Adam Toon. Interpreting Quantum Theory Simon Friederich. Scientific Enquiry and Natural Kinds P.