Institute of

Insights Advanced
Study

Models: Parables v Fables

Nancy Cartwright

Volume 1 2008 Number 11

ISSN 1756-2074
Institute of Advanced Study Insights

About Insights
Insights captures the ideas and work-in-progress of the Fellows of the Institute of Advanced Study
at Durham University. Up to twenty distinguished and ‘fast-track’ Fellows reside at the IAS in any
academic year. They are world-class scholars who come to Durham to participate in a variety of events
around a core inter-disciplinary theme, which changes from year to year. Each theme inspires a new
series of Insights, and these are listed in the inside back cover of each issue. These short papers take
the form of thought experiments, summaries of research findings, theoretical statements, original
reviews, and occasionally more fully worked treatises. Every fellow who visits the IAS is asked to write
for this series. The Directors of the IAS – Ash Amin, Michael O’Neill, Susan J. Smith and Colin Bain –
also invite submissions from others involved in the themes, events and activities of the IAS.

About the Institute of Advanced Study

The Institute of Advanced Study, launched in October 2006 to commemorate Durham University’s 175th
Anniversary, is a flagship project reaffirming the value of ideas and the public role of universities. The
Institute aims to cultivate new thinking on ideas that might change the world, through unconstrained
dialogue between the disciplines as well as interaction between scholars, intellectuals and public
figures of world standing from a variety of backgrounds and countries. The Durham IAS is one of only
a handful of comparable institutions in the world that incorporates the Sciences, Social Sciences, the
Arts and the Humanities.

The focal point of the IAS is a programme of work associated with, but not exclusive to, an annual
research theme. At the core of this work lies a prestigious Fellowship programme. This programme
gathers together scholars, intellectuals and public figures of world standing or world-promise to address
topics of major academic or public interest. Their mission is to anticipate the new and re-interpret the
old, communicating across and working between disciplinary boundaries.

Every year, the Institute invites as many as twenty highly creative individuals to spend up to three
months in Durham. They are located in Cosin’s Hall, a magnificent and spacious 18th century mansion
which, together with Durham Cathedral and Durham Castle, forms part of Palace Green, dominating
the World Heritage Site of Durham Peninsula. During their stay, Fellows engage with departments
and colleges, deliver public lectures and seminars, and, above all, join an international community of
researchers to address the theme selected for that year. Further details of the IAS and its Fellowship
programme can be found at www.durham.ac.uk/ias/fellows

Copyright
The design and contents of Insights are subject to copyright. Copyright and Reproduction Rights in all
submitted contributions remain with the authors, as described in the Author’s Copyright Agreement.
Copyright and Reproduction Rights of all other material remain with Insights.

Except under the terms of Fair Dealing (UK Copyright, Designs and Patents Act 1988), the user may
not modify, copy, reproduce, retransmit or otherwise distribute the site and its contents (whether
text, graphics or original research concepts), without express permission in writing from the Institute.
Where the above content is directly or indirectly reproduced in an academic context under the terms
of Fair Dealing, this must be acknowledged with the appropriate bibliographical citation.

The opinions stated in the Insights papers are those of their respective authors and do not necessarily
reflect the opinions of the Institute of Advanced Study, Durham University, or the staff and students
thereof.
Institute of Advanced Study Insights

MODELS: PARABLES V FABLES

A good many models used in physics and economics offer descriptions

of imaginary situations, using a combination of mathematics and natural
language. The descriptions are both thin – not much about the situation
is filled in – and unrealistic – what is filled in is not true of many real
situations. Yet we want to use the results of these models to inform our
conclusions about a range of actually occurring situations. I propose we
interpret many of these models as fables. The actual happenings in the
model are only one concrete instantiation of a lesson that is more widely
applicable ‘when expressed in more abstract language.’ A fable generally
comes with the abstract lesson attached; parables, by contrast, usually
require an interpretation from outside. Models, then, tend to be more like
parables than fables. The right abstract interpretation is not supplied by
the model itself. The central problem, then, is where shall it come from?

The Problem of Unrealistic Assumptions, Mach 1:

Valid Arguments but False Premises

T he models I shall discuss here, found typically in physics and economics, offer descriptions
of imaginary situations or systems using a combination of mathematics and natural
language. The descriptions are thin: not much about the situation is filled in. They are often
unrealistic as well in that what is filled in is not true of many real situations. Yet in many
cases we want to use the results of these models to inform our conclusions about a range of
actually occurring (so-called target) situations.

I am also going to restrict my attention to models in which results are derived by deduction.
The whole point of these formal models is rigour, which is why they are preferred by physicists
and economists alike over more informal reasonings that merely make results plausible.
Deduction is a key ingredient in this rigour. We are assured that the consequences drawn
from the models are genuine because they follow deductively from the starting descriptions;
these consequences must occur whenever these descriptions are satisfied.

The ‘unrealistic’ assumptions that are offered in a model’s descriptions are no problem so
long as they play no role in deducing the intended results of the model. But this is seldom
the case: in fact quite the contrary. They are often necessary to the deductions offered in the
model.¹ This gives rise to the canonical ‘problem of unrealistic assumptions’: how can a result
that must occur given characteristics different from those in the target inform conclusions
about what will happen in the target? The conclusion is supposed to be guaranteed because
it follows deductively from the premises. How does that provide information about what
conclusions to expect when the premises are different?

2
Institute of Advanced Study Insights

The Plan

I n tackling the problem of unrealistic assumptions this paper will weave together three
different strands of enquiry:
• previous work of mine on Galilean thought experiments;
• previous work of mine on models as fables;
• Menno Rol’s insight that abstraction can turn falsehoods into truths.

It will also wind through an increasingly narrow spiral:

• from ‘problem, Mach 1’: the first broad problem of unrealistic assumptions just described;
• to ‘solution, Mach 1’ that should work for certain specific kinds of models – those that can
count as Galilean thought experiments;
• to a twist at which the problem re-emerges in narrower form, as ‘problem, Mach 2’: the
problem of overconstraint;
• another twist provided by Menno Rol offers ‘solution, Mach 3’ to this new version of the
problem for an even more restricted set of models – by abstraction of the kind seen in the
moral of a fable;
• finally ending in a third version, ‘problem, Mach 3’ of the problem of unrealistic assumptions,
which arises because models are far more like parables than fables.

This final version of the problem of unrealistic assumptions cannot be solved within the
model itself nor by philosophy. The problem in the end demands that the model be located
in a strong, rich scientific network that can pick out the right abstract concepts with which to
formulate the model’s results.

Strand 1. Solution, Mach 1: Galilean Thought Experiments

U nrealistic assumptions do not always stand in the way of drawing lessons about real
situations from models. Some models function as Galilean thought experiments and for
these, unrealistic assumptions are not a hindrance but a necessity. A Galilean experiment, as I
use the term, isolates a single factor to observe its natural effect when it operates ‘on its own’
and ‘without impediment.’ In a real Galilean experiment the effect is produced in accord with
the laws of nature. In a Galilean thought experiment it is the principles built into the model
that determine what the effect must be. So, real experiments and thought experiments have
complementary virtues. In the real experiment we can never be sure that we have eliminated
all confounding factors but we can be sure the effect is produced in accord with nature’s laws.
By contrast the situation described in the thought experiment has only the factors in it that
we put there. So we can be sure that confounders are absent but we cannot be sure the effect
is right because that depends on the principles we provide in the model.

Typical economics models, and many in physics as well – especially those set as problems to
work out in physics texts – can certainly be taken to be Galilean thought experiments, isolating
a single factor to study what it does ‘on its own.’ This is clear not only from the practice in
both cases; it is explicit in much economics discussion and in some from physics as well. For
those models that do serve as Galilean thought experiments the unrealistic assumptions that
suppose the factor is at work all on its own, with no confounders and no impediments, are no
more of a problem than they are for real Galilean experiments. If we can learn about target
situations with more ‘realistic’ arrangements from actual Galilean experiments, despite the
‘unrealistic’ assumptions necessary to the experiment, the same is true for Galilean thought

3
Institute of Advanced Study Insights

experiments, so long as the basic principles used in the model to drive the consequences are
accurate enough.² So at least for some models and some kinds of unrealistic assumptions,
unrealistic assumptions pose no problem.

The Problem of Unrealistic Assumptions, Mach 2: Overconstraint

T his is a rather too happy conclusion however. That is because a good many of the models
that can be cast as Galilean thought experiments have a number of ‘unrealistic’ assumptions
beyond those necessary for them to count as Galilean experiments – that is beyond those that
eliminate confounders and impediments. This is generally for two interacting reasons.

First, causes cannot just act in a vacuum. They need a concrete setting in which to play
out. Consider for example a model to study the effect of skill loss during unemployment on
future employment levels. If it is to be a Galilean model there must be no further source of
unemployment at work in it: no downturns in consumer spending, no shift to a war economy,
and so forth; no motives that differ between when employers invest to open future jobs
from when they do not, other than the difference in profit they expect due to a difference in
the efficiency of the workers. Still, the model needs employers in it in order to study what
happens, given their different expectations; and it needs to have workers who have lost skills
and workers who have not to create these different expectations. How many workers, how many
employers, what is the ratio of employed to unemployed, etc? These factors are not properly
thought of as causes or impediments to the mechanism by which skill loss might affect future
employment levels. Hence the answer to what form they should take is not dictated by the
demand of the Galilean experiment. Still they must take some form or other, otherwise skill
loss cannot be set operating.

Second, is the well-rehearsed reason that matters must often be set in very particular ways if
calculations and deductions are to be at all possible. So often, mathematically more tractable
descriptions are substituted for descriptions that are more true to the target situations that
we want the model results to bear on. Indeed it is often the need for mathematical tractability
that solves the first problem by settling how to fix the concrete setting in which the isolated
cause will play out.

So, most Galilean thought experiments have many more ‘unrealistic’ assumptions than they
should. Again, this would not be a problem if these assumptions did not play a role in
deducing the final results. But of course generally they do – that is the point of including
them in the first place. Just by inspection we can see that they are a necessary part of the
deduction offered by the model.3

In these cases I say that the results of the model are overconstrained. All the conditions
sufficient to ensure that the model describes a Galilean experiment are met. So (pace mistakes
in the driving principles) the results must be ones we would see in a real Galilean experiment.
The problem is that the Galilean experiment takes place in a very special and unusual setting.
What we see is indeed the result of the cause acting on its own without impediment but it is
a very special result that we cannot expect in other Galilean experiments. We know we cannot
expect it because we can see by inspection that the description of the special setting plays
a necessary role in the derivation offered. So, unrealistic assumptions that overconstrain
the results are a problem for learning lessons that apply elsewhere even if the model does
function as a Galilean thought experiment.4

4
Institute of Advanced Study Insights

In order to explain the proposed solution to this new problem of overconstraint, I first turn to
another topic altogether, that of fables and their morals. I shall spend quite a bit of time on
this topic because doing so will make it easy to see Rol’s proposal, which I summarize in the
slogan, ‘Abstraction can turn falsehoods into truths.’

Strand 2: Fables, Models and Morals

S o far I have discussed models that are used to describe thought experiments; these are
models that look to the world. Let us turn now to a different kind of model, one that faces
theory: models I label as ‘interpretive.’ According to positivist accounts theoretical principles
come in two breeds: internal principles, which lay out relations among theoretical concepts,
and bridge principles, which link theoretical concepts with ones that we already know how to
apply to the world. The latter are important for the predictive power of theory: there are rules
within the theory – language entry rules – for how and when to apply its concepts so that we
can then make use of the links within the theory to make predictions about real situations.
These matter especially with respect to abstract mathematical descriptions and physics is rife
with them. Consider the following, which are just two examples from a very long list of similar
well-known principles.
• Two bodies of mass m1 and m2 separated by a distance r can be assigned a force between
them of size Gm1m2/r2
• Two bodies of charge q1 and q2 separated by a distance r can be assigned a force between
them of size eq1q2/r2

The left-hand side of these bridge principles are descriptions that I label interpretive models.
The right-hand side is a mathematical representation from classical gravitational and
electromagnetic theories. So bridge principles link interpretive models with the theoretical
terms that can be assigned to a situation when those interpretive models apply to it.

The interpretive model, I have argued, is like a fable and the theoretical description it licenses
is its moral. I do so in order to stress that the relationship between theoretical description and
the description in the model that licenses it is that of the abstract to the concrete, following
the theory of the fable defended by Gotthold Ephraim Lessing, the great critic and dramatist
of the German Enlightenment.5

Lessing argues, ‘In order to give a general symbolic conclusion all the clarity of which it
is capable, that is in order to elucidate it as much as possible, we must reduce it to the
particular in order to know it intuitively’ (Lessing, 1759 [1967], p. 100). For him this is not
just a matter of anschaulichkeit – intuitive understanding. It is also a matter of ontology: ‘The
general exists only in the particular….’ (Lessing, 1759 [1967], p. 73). This is the aspect I
want to stress about theoretical terms that get applied via bridge principles.

Some theoretical terms do not have bridge principles; they apply to the world ‘directly.’ For
these we may have a variety of different ways to measure or to test when they apply but
there is no other further description that must be met to license their application. Bridge
principles then have a dual effect. On the one hand they provide more visualizable content to
theoretical terms and an aid to their application; on the other they place a strong constraint
on the domain of these terms. As Lessing says, the general exists only in the particular. These
abstract theoretical terms apply only given the applicability of their more concrete models.
These constraints, however, are a large part of the reason why hypothetico-deductive testing
in physics has so much confirmatory power, unlike in other fields where we are inclined to

5
Institute of Advanced Study Insights

think of the explanations offered as ‘just so’ stories with little power to confirm the theory that
provides them.

I illustrate the relation of the abstract to the concrete, following Lessing, with a fable of his
own:
A marten eats the grouse;
A fox throttles the marten; the tooth of the wolf, the fox.
Moral: the weaker are always prey to the stronger.
As I described in The Dappled World (Cartwright, 1999), Lessing makes up this story as part
of his argument to show that a fable is no allegory. Allegories say not what their words seem
to say, but rather something similar. But where is the allegory in the fable of the grouse, the
marten, the fox and the wolf: ‘What similarity here does the grouse have with the weakest, the
marten with the weak and so forth? Similarity! Does the fox merely resemble the strong and
the wolf the strongest or is the former the strong, the latter the strongest? He is it’ (Lessing,
1759 [1967], p. 73). For Lessing similarity is the wrong idea to focus on. The relationship
between the moral and the fable is that of the general to the more specific and it is ‘a kind of
misusage of the word to say that the special has a similarity with the general, the individual
with its type, the type with its kind’ (Lessing, 1759 [1967], p. 73). Each particular is a case
of the general under which it falls.

The point comes up again when Lessing protests against those who maintain that the moral is
hiding in the fable or disguised there. Lessing argues: ‘How can one disguise (verkleiden) the
general in the particular… If one insists on a similar word here it must at least be einkleiden
rather than verkleiden.’ Einkleiden is to fit out, as when you take the children to the shops in
the autumn to buy them new school clothes. So the moral is to be ‘fitted out’ by the fable.

The account of abstraction that I borrow from Lessing provides two necessary conditions:
• First, a concept that is abstract relative to another more concrete set of descriptions
never applies unless one of the more concrete descriptions also applies. These are the
descriptions that can be used to fit out the abstract description on any given occasion.
• Second, satisfying the associated concrete description that applies on a particular
occasion is what satisfying the abstract description consists in on that occasion.
This discussion of Lessing began with the topic of interpretive models. My topic in this paper
is not interpretive models, however, which are models that link to theory, but rather unrealistic
models that somehow, despite their unrealistic assumptions, link to the world. What I want
to take away for the current study from the earlier work on interpretive models is the idea of
how the model relates to the lesson to be drawn from it. Like fables and their morals, the
lesson of the model is abstract relative to the more concrete descriptions of the model. Like
the fable, the model ‘fits out’ the more abstract lesson; and when a situation satisfies the
more concrete description from the model, that is what it is for that situation to satisfy the
more abstract lesson.

Strand 3. Solution, Mach 3: From Falsehood to Truth via Abstraction

T he problem of models with unrealistic assumptions is one of the standard worries both in
the philosophy of economics and in economics itself. Philosopher of economics Menno
Rol has a nice account of why it need not always be a problem. One can, he argues, go from
falsehood to truth by climbing up the ladder of abstraction (Rol, 2008). Rather than delving
into economics, let me illustrate his point with a physics example that I think will be familiar
to everyone.

6
Institute of Advanced Study Insights

Suppose we perform a careful Galilean experiment to see how bodies move inertially, that
is, subject to no forces. We do it perfectly; we succeed in stripping (or calculating) away all
forces. But we do our experiments on a Euclidean plane. From this we conclude that bodies
moving inertially will follow a Euclidean straight line. This conclusion is entirely correct in
the setting of the experiment. But it need not be true elsewhere. In particular this will not
describe correctly inertial motion in a spherical geometry, where the body subject to no forces
will move on a great circle. To use my earlier language, we succeed in carrying out an ideal
Galilean experiment but the results are overconstrained. The solution, following Rol, is to
describe the results of the experiment equally correctly in more abstract vocabulary: The
bodies in the experiment travel on geodesics – that is, they take the shortest distance between
any two points in the relevant geometry. This conclusion is true both in the experiment we
conduct and (putatively) everywhere else as well.

This account dovetails exactly with the image of models as fables. The lesson of the model is,
properly, more abstract than what is seen to happen in the model and that can be described in
the concepts introduced there. In the model the marten eats the grouse; the body moves along
a Euclidean line. The lesson is that the weaker are prey to the stronger; that inertial bodies
move on geodesics. The abstract lesson can be true of a variety of new, different situations
where the more concrete behaviour will fail.

The advantage of thinking of what happens here in terms of Lessing’s account of morals and
fables is that it makes clear that there is nothing wrong with the initial experiment. What
is wrong vis-à-vis applicability elsewhere is the level at which the conclusion is described.
Moreover, no experiment could have done better. Experiments must be performed in some
geometry or other. That is the point of invoking Lessing’s theory of the relation of the abstract
to the concrete. The abstract can exist only in the concrete. You cannot get it unless it is
fitted out in one way or another. What the abstract consists in given one filling out will be very
different from what it consists in given another. For the marten and the grouse, the grouse’s
being weaker consists in being slow and not having sharp teeth, claws or a hard shell; being
prey is being eaten. For a worker vis-à-vis employer, being weaker can consist in having no
union, no transferable in-demand skills and no wealth; and being prey to equals working long
hours in bad conditions for little pay. Still, both are cases where the weaker are prey to the
stronger. And in any case, it cannot just be true that someone is weaker and prey to another.
In every case there must be something more concrete – and thus less generalizable – that this
consists in.

My topic here is thought experiments not real experiments. But the same lessons apply. A
thought experiment can succeed perfectly in isolating the factor under study and observing –
correctly – what it does on its own, without impediment. But if the results are overconstrained
they will not readily generalize. Yet, just as with the ‘real’ Galilean experiment I described for
inertia, there may be no alternative. The experiment must be performed in some geometry or
other. Similarly, the model to study the effects of skill loss during unemployment on future
unemployment rates may have only two generations of workers and one employer, where these
affect the outcome though they could not properly be counted as impediments to skill loss
in affecting future unemployment nor as confounding causes. But all situations have some
generational structure among the workers and some number of employers. ‘Real’ economic
experiments cannot eliminate them either and so will also be overconstrained.

Thinking of thought experiments as fables, then, points out two important methodological
lessons:

7
Institute of Advanced Study Insights

• Though the results of an experiment or of a thought experiment may be overconstrained, this

may be inevitable since the abstract exists only in the concrete.
• To get a conclusion that is true both in the model and in a variety of other cases, it may well
be necessary to follow Rol’s advice and climb up the ladder of abstraction.

Strand 3. The Problem of Unrealistic Assumptions, Mach 3: Not Fables

but Parables

C onsider the parable of the prodigal son or of the good Samaritan. Why are these parables
and not fables? One reason is that they are open to interpretation. The moral is not written
in but must be supplied from elsewhere. Defending a moral as the correct one requires a
great deal of outside work, including much interpretation of other parts of the available text
and of the world itself and how it operates.

So, too, with our ‘unrealistic’ models. Many of these may be Galilean thought experiments
and so rightly have ‘unrealistic’ assumptions. And in many cases the correct lessons to be
drawn may be more abstract than those described immediately in the concrete situation of the
model. But seldom can we really cast the models as fables because the moral is not written
in. They are rather like parables, where the prescription for drawing the right lesson must
come from elsewhere. Theory can help here, as can a wealth of other cases to look to, and
having a good set of well-understood more abstract concepts to hand will play a big role. So
the good news that one can move from falsehood in a model to truth by climbing up the ladder
of abstraction is considerably dampened by the fact that the model generally does not tell us
which ladder of abstraction to use and how far to climb.

I should stress that this problem is not peculiar to thought experiments. As I have mentioned,
‘real’ experiments can be overconstrained too. As with thought experiments this need not be
a problem since, as with fables and their morals, what results in the (correctly conducted)
overconstrained experiment will be what the generalizable result consists in for that situation;
it will be an instance of the generalizable conclusion. But the experiment does not show
what the generalizable conclusion actually is, how far up which ladder of abstraction one
must climb to reach a result that will be true of new target situations as well. This is, I
think, clearly recognized in physics and in much of economics as well, even though not
articulated in this way. I stress it because I think that it has not been taken on board in the
new drive for experiments in evidence-based policy, where practitioners are trying to draw
general conclusions without the aid of theory or appeal to a set of well-understood abstract
concepts whose reliability has been established elsewhere. So it is important to stress that
real experiments, just like thought experiments, are far more often parables than fables.

Conclusion

I f we are to use Galilean thought experiments to inform ourselves about target situations we
had better recognize that these models are more like parables than fables. So constructing
the model and deriving its consequences are just a small step towards drawing a lesson from
it. In order to know what the parable means we need to study a great deal of text, reading
both the theory that imbeds the model and reading the world itself.

8
Institute of Advanced Study Insights

Acknowledgement

I am very grateful for the support I received at the Institute of Advanced Study at the
University of Durham whilst carrying out this work.

Notes

¹ The requisite deduction will sometimes not be literally on offer in the model but rather
presumed.

² In either case exporting from the Galilean experiment requires both more and stronger
assumptions than those supplied in the experiment. My own view is that exporting often
employs the logic of capacities, where the assumption that a factor has a capacity to study
in the first place takes a great deal of highly varied independent evidence. (Cf. Nancy
Cartwright’s Nature’s Capacities and their Measurement (1989) and ‘What is this Thing
Called “Efficacy”?’ (forthcoming, 2009).)

3
Explicit attempts to deal with this problem often involve so-called ‘robustness’
investigations: vary these extra assumptions in different ways to see if the results are
still more or less the same. Then, I suppose, we are meant to do a quick induction to the
conclusion that the results will be the same under the conditions that hold in the target
situations. Not only is this inductive inference extremely dicey but usually the variation
is not very great. Also, often the interest is not so much in varying the ‘extra-Galilean’
assumptions but rather in adding in some further causes to see how the results are affected
when a more realistic arrangement of causes occurs. This latter offers some help with the
problem of whether the results are exportable from the experiment to other situations – the
question ‘Can an induction be done at all?’ – but not with the problem of which results to
export.

4
For a more detailed description of Galilean thought experiments and the problem of
over constraint see Nancy Cartwright, ‘The Vanity of Rigour in Economics,’ chapter 15 in
Hunting Causes and Using Them (2007).

5
My discussion of Lessing here is taken directly from The Dappled World, 1999, pp. 35-
48.

9
Institute of Advanced Study Insights

Reference List

Cartwright, N. D. (1989) Nature’s Capacities and Their Measurement. Oxford: Oxford

University Press.

- - - . (1999) The Dappled World. A Study of the Boundaries of Science. Cambridge:

Cambridge University Press.

- - - . (2007) Hunting Causes and Using Them. Approaches in Philosophy and Economics.
Cambridge: Cambridge University Press.

- - - . (Forthcoming, 2009) What is this Thing Called ‘Efficacy’? In Mantzavinos, C. (ed.)

Philosophy of the Social Sciences. Philosophical Theory and Scientific Practice. Cambridge:
Cambridge University Press.

Lessing, G. E. (1759 [1967]) Abhandlungen über die Fable. Stuttgart: Philipp Reclam.

Rol, M. (2008) Idealization, abstraction, and the policy relevance of economic theories.
Journal of Economic Methodology 15-1: 69-97.

10
Institute of Advanced Study Insights

Backlist of Papers Published in Insights

2008 Volume 1

No. Author Title Series

1 Boris Wiseman Lévi-Strauss, Caduveo Body Painting General

and the Readymade: Thinking Borderlines

2 John Hedley Brooke Can Scientific Discovery be a Darwin’s Legacy

Religious Experience?

3 Bryan R. Cullen Rapid and Ongoing Darwinian Darwin’s Legacy

Selection of the Human Genome

4 Penelope Deutscher Women, Animality, Immunity – and Darwin’s Legacy

the Slave of the Slave

5 Martin Harwit The Growth of Astrophysical Modelling

Understanding

6 Donald MacKenzie Making Things the Same: Gases, Modelling

Emission Rights and the Politics of
Carbon Markets

7 Lorraine Code Thinking Ecologically about Biology Darwin’s Legacy

8 Eric Winsberg A Function for Fictions: Expanding the Modelling

Scope of Science

9 Willard Bohn Visual Poetry in France after Apollinaire Modelling

10 Robert A. Skipper Jr R. A. Fisher and the Origins of Random Darwin’s Legacy

Drift

Insights

Insights is edited by Susan J. Smith, IAS Director and Professor of Geography.

Correspondence should be directed to Audrey Bowron ([email protected]).