Full Chapter Model Based Reasoning in Science and Technology Logical Epistemological and Cognitive Issues 1St Edition Lorenzo Magnani PDF
Full Chapter Model Based Reasoning in Science and Technology Logical Epistemological and Cognitive Issues 1St Edition Lorenzo Magnani PDF
Full Chapter Model Based Reasoning in Science and Technology Logical Epistemological and Cognitive Issues 1St Edition Lorenzo Magnani PDF
https://textbookfull.com/product/springer-handbook-of-model-
based-science-1st-edition-lorenzo-magnani/
https://textbookfull.com/product/logical-studies-of-
paraconsistent-reasoning-in-science-and-mathematics-1st-edition-
holger-andreas/
https://textbookfull.com/product/a-science-based-critique-of-
epistemological-naturalism-in-quines-tradition-reto-gubelmann/
https://textbookfull.com/product/logical-reasoning-and-data-
interpretation-for-cat-6th-edition-nishit-k-sinha/
Communicating Science and Technology in Society: Issues
of Public Accountability and Engagement Ana Delicado
https://textbookfull.com/product/communicating-science-and-
technology-in-society-issues-of-public-accountability-and-
engagement-ana-delicado/
https://textbookfull.com/product/doing-philosophy-from-common-
curiosity-to-logical-reasoning-timothy-williamson/
https://textbookfull.com/product/how-to-prepare-for-logical-
reasoning-for-cat-arun-sharma-sharma/
https://textbookfull.com/product/tantalum-and-niobium-based-
capacitors-science-technology-and-applications-1st-edition-yuri-
freeman-auth/
https://textbookfull.com/product/cognitive-reasoning-for-
compliant-robot-manipulation-daniel-sebastian-leidner/
Studies in Applied Philosophy,
Epistemology and Rational Ethics
Lorenzo Magnani
Claudia Casadio Editors
Model-Based
Reasoning in
Science and
Technology
Logical, Epistemological, and Cognitive
Issues
Studies in Applied Philosophy, Epistemology
and Rational Ethics
Volume 27
Series editor
Lorenzo Magnani, University of Pavia, Pavia, Italy
e-mail: [email protected]
Editorial Board
Atocha Aliseda
Universidad Nacional Autónoma de México (UNAM), Coyoacan, Mexico
Giuseppe Longo
Centre Cavaillès, CNRS—Ecole Normale Supérieure, Paris, France
Chris Sinha
Lund University, Lund, Sweden
Paul Thagard
Waterloo University, Waterloo, ON, Canada
John Woods
University of British Columbia, Vancouver, BC, Canada
About this Series
Advisory Board
Editors
Model-Based Reasoning
in Science and Technology
Logical, Epistemological, and Cognitive
Issues
123
Editors
Lorenzo Magnani Claudia Casadio
Department of Humanities, Philosophy Department of Philosophy, Education and
Section, and Computational Philosophy Economical-Quantitative Sciences
Laboratory University of Chieti and Pescara
University of Pavia Chieti
Pavia Italy
Italy
This volume is a collection of selected papers that were presented at the interna-
tional conference Model-Based Reasoning in Science and Technology. Models and
Inferences: Logical, Epistemological, and Cognitive Issues (MBR015_ITALY),
held at the Centro Congressi Mediaterraneo, Sestri Levante, Italy, June 25–27,
2015, chaired by Lorenzo Magnani.
A previous volume, Model-Based Reasoning in Scientific Discovery, edited by
L. Magnani, N.J. Nersessian, and P. Thagard (Kluwer Academic/Plenum
Publishers, New York, 1999; Chinese edition, China Science and Technology
Press, Beijing, 2000), was based on the papers presented at the first “model-based
reasoning” international conference, held at the University of Pavia, Pavia, Italy, in
December 1998. Other two volumes were based on the papers presented at the
second “model-based reasoning” international conference, held at the same place in
May 2001: Model-Based Reasoning. Scientific Discovery, Technological
Innovation, Values, edited by L. Magnani and N.J. Nersessian (Kluwer
Academic/Plenum Publishers, New York, 2002), and Logical and Computational
Aspects of Model-Based Reasoning, edited by L. Magnani, N.J. Nersessian, and C.
Pizzi (Kluwer Academic, Dordrecht, 2002). Another volume, Model-Based
Reasoning in Science and Engineering, edited by L. Magnani (College
Publications, London, 2006), was based on the papers presented at the third
“model-based reasoning” international conference, held at the same place in
December 2004. The volume Model-Based Reasoning in Science and Medicine,
edited by L. Magnani and P. Li (Springer, Heidelberg/Berlin 2006), was based on
the papers presented at the fourth “model-based reasoning” conference, held at Sun
Yat-sen University, Guangzhou, P.R. China. The volume Model-Based Reasoning
in Science and Technology. Abduction, Logic, and Computational Discovery,
edited by L. Magnani, W. Carnielli and C. Pizzi (Springer, Heidelberg/Berlin
2010), was based on the papers presented at the fifth “model-based reasoning”
conference, held at the University of Campinas, Campinas, Brazil, in December
2009. Finally, the volume Model-Based Reasoning in Science and Technology.
Theoretical and Cognitive Issues, edited by L. Magnani, (Springer,
v
vi Preface
support is gratefully acknowledged. The preparation of the volume would not have
been possible without the contribution of resources and facilities of the
Computational Philosophy Laboratory and of the Department of Humanities,
Philosophy Section, University of Pavia.
Several papers concerning model-based reasoning deriving from the previous
conferences MBR98 and MBR01 can be found in special issues of Journals: in
Philosophica: Abduction and Scientific Discovery, 61(1), 1998, and Analogy and
Mental Modeling in Scientific Discovery, 61(2) 1998; in Foundations of Science:
Model-Based Reasoning in Science: Learning and Discovery, 5(2) 2000, all edited
by L. Magnani, N.J. Nersessian, and P. Thagard; in Foundations of Science:
Abductive Reasoning in Science, 9, 2004, and Model-Based Reasoning: Visual,
Analogical, Simulative, 10, 2005; in Mind and Society: Scientific Discovery:
Model-Based Reasoning, 5(3), 2002, and Commonsense and Scientific Reasoning,
4(2), 2001, all edited by L. Magnani and N.J. Nersessian. Finally, other related
philosophical, epistemological, and cognitive-oriented papers deriving from the
presentations given at the conference MBR04 have been published in a special
issue of the Logic Journal of the IGPL: Abduction, Practical Reasoning, and
Creative Inferences in Science, 14(1) (2006), and have been published in two
special issues of Foundations of Science: Tracking Irrational Sets: Science,
Technology, Ethics, and Model-Based Reasoning in Science and Engineering, 13
(1) and 13(2) (2008), all edited by L. Magnani. Other technical logical papers
presented at MBR09_BRAZIL have been published in a special issue of the Logic
Journal of the IGPL: Formal Representations in Model-Based Reasoning and
Abduction, 20(2) (2012), edited by L. Magnani, W. Carnielli, and C. Pizzi. Finally,
technical logical papers presented at MBR12_ITALY have been published in a
special issue of the Logic Journal of the IGPL: Formal Representations in
Model-Based Reasoning and Abduction, 21(6) (2013), edited by L. Magnani.
Other more technical formal papers presented at (MBR015_ITALY) will be
published in a special issue of the Logic Journal of the IGPL, edited by L. Magnani
and C. Casadio.
ix
x Contents
Otávio Bueno
Abstract Diagrams are hybrid entities, which incorporate both linguistic and
pictorial elements, and are crucial to any account of scientific and mathematical
reasoning. Hence, they offer a rich source of examples to examine the relation
between model-theoretic considerations (central to a model-based approach) and
linguistic features (crucial to a language-based view of scientific and mathematical
reasoning). Diagrams also play different roles in different fields. In scientific
practice, their role tends not to be evidential in nature, and includes: (i) highlighting
relevant relations in a micrograph (by making salient certain bits of information);
(ii) sketching the plan for an experiment; and (iii) expressing expected visually
salient information about the outcome of an experiment. None of these traits are
evidential; rather they are all pragmatic. In contrast, in mathematical practice,
diagrams are used as (i) heuristic tools in proof construction (including dynamic
diagrams involved in computer visualization); (ii) notational devices; and
(iii) full-blown proof procedures (Giaquinto 2005; and Brown in Philosophy of
mathematics. Routledge, New York, 2008). Some of these traits are evidential.
After assessing these different roles, I explain why diagrams are used in the way
they are in these two fields. The result leads to an account of different styles of
scientific reasoning within a broadly model-based conception.
1 Introduction
The sematic view of theories emphasizes the role played by models in scientific
practice, and it tends to downplay the corresponding role for linguistic considera-
tions. As Bas van Fraassen, one of the major advocates of that view, points out:
The syntactic picture of a theory identifies it with a body of theorems, stated in one
particular language chosen for the expression of that theory. This should be contrasted with
O. Bueno (&)
Department of Philosophy, University of Miami, Coral Gables,
FL 33124-4670, USA
e-mail: [email protected]
the alternative of presenting a theory in the first instance by identifying a class of structures
as its models. In this second, semantic, approach the language used to express the theory is
neither basic nor unique; the same class of structures could well be described in radically
different ways, each with its own limitations. The models occupy centre stage. (van
Fraassen 1980, p. 44)
This puts pressure on the semantic view to accommodate those aspects of scientific
practice that rely on linguistic considerations, such as various styles of scientific
reasoning.
In contrast, the received view, in the hands of Rudolf Carnap, for instance,
emphasizes the importance of linguistic considerations for the proper understanding
of science.
Apart from the questions of the individual sciences, only the questions of the logical
analysis of science, of its sentences, terms, concepts, theories, etc., are left as genuine
scientific questions. We shall call this complex of questions the logic of science. […]
According to this view, then, once philosophy is purified of all unscientific elements, only
the logic of science remains. […] [T]he logic of science takes the place of the inextricable
tangle of problems which is known as philosophy (Carnap 1934/1937, p. 279).
1
For a survey of the semantic and the received views, see Suppe (1977a, b) and references therein.
Visual Reasoning in Science and Mathematics 5
The semantic and the received views can be compared under many dimensions.
For the purposes of this paper, I will focus on only two of them: the role of
linguistic features in the proper understanding of central aspects of scientific and
mathematical practice, and the role played by models—broadly understood to
include diagrams—in this practice. I will also consider strong formulations of these
views: for the semantic approach (on this strong reading), linguistic considerations
are largely irrelevant, and what are crucial are the relevant models, whereas for the
received view (also on a strong reading), linguistic considerations—including, in
particular, the requirement of formalization—are crucial, whereas models are not so
central.
Diagrams, I just noted, play multiple roles in scientific practice. I will start by
providing some examples to illustrate these roles. The examples, of course, are not
meant to be comprehensive, but they highlight significant roles played by diagrams
in this context.
(i) A diagram may express expected visually salient information about the
outcome of an experiment. DNA nanotechnology involves the use of DNA as
a biomimetic component for self-assembly (Seeman 2003, 2005; Seeman and
Belcher 2002). One interesting experiment involved the construction of
certain arrangements of DNA strands in predetermined shapes (Ding et al.
2004). First, a triangular arrangement was designed, followed by a hexagon
formed by six DNA triangles suitably positioned. Finally, with multiple such
DNA hexagons properly arranged, a DNA honeycomb is formed. A diagram
is designed to express what researchers expect to detect with the output of the
experiment (see Fig. 1, on the left). The outcome of the experiment, which
was conducted with an atomic force microscope (AFM), seems to support the
intended result, as honeycomb structures can be “seen” on the AFM image
(see Fig. 2, on the right). (For additional discussion of this case, see Bueno
2011.)
It is important to note that what provides evidence for the intended con-
clusion, in this instance, is the AFM image rather than the diagram, which is
only a representation of what the researchers expected to detect when the
experiment was conducted.
(ii) A diagram may highlight relevant information, for instance, on a micro-
graph. Sometimes a diagram is drawn on a micrograph to make salient
certain bits of information. This is seen in Fig. 1 (on the right). Consider the
bottom right corner micrograph: in order to highlight the intended hexagon, a
diagram of this shape is drawn on the AFM image. As a result, the hexagon
is made salient (for further discussion, see Lynch 1991; Bueno 2011).
6 O. Bueno
Fig. 1 The diagram of the DNA strand experiment (left) and, the corresponding AFM image
(right) (Ding et al. 2004, p. 10230)
(iii) Diagrams can also be used to sketch the plan of an experiment. L.A. Bumm
and a group of collaborators once tried to construct a single-molecular wire by
establishing a conducting single-molecular current through a non-conducting
medium (Bumm et al. 1996). A diagram was created to sketch the plan of the
experiment (see Fig. 2, left). After dropping a conducting material through a
non-conducting medium, researchers expected that single-molecular wires
would be formed. The tip of the scanning-tunneling microscope (STM) would
then establish a current through the non-conducting material via each ‘single’-
molecular wire all the way to the gold substratum. The STM micrograph
exhibits the ‘single’-molecular wires, viewed from the top: the little white
blobs on the micrograph (Fig. 2, right) indicate ‘single’-molecular wires; big
white blobs indicate multiple-molecular wires.
In the end, however, it was unclear that just a single molecule was involved,
rather than, say, a couple of molecules, given that a virtually indistinguishable STM
micrograph would have been produced in each case. In this case, the evidence
didn’t support the conclusion, since it was unable to rule out possibilities that
undermine the conclusion (for additional discussion, see Bueno 2011).
Visual Reasoning in Science and Mathematics 7
Fig. 2 ‘Single’-molecular wire experiment: diagrammatic sketch (left) and STM output (right)
(Bumm et al. 1996)
Fig. 3 Mathematical
diagrams provide
explanations of certain
mathematical relation
Visual Reasoning in Science and Mathematics 9
(ii) Mathematical diagrams can also provide heuristic tools in proof construction.
Consider, again, Euler’s formula. It is clear that if we expand or contract the
coordinates of the eih vector by a real magnitude r (that is, ‘r’ stands for a real
number), so that we have rcos h and rsin h, the corresponding vector will
expand or contract by a factor r. We thus have the geometrical significance of
the following simple result (Giaquinto 2005, pp. 79–80):
This result is clearly expressed in the diagram (see Fig. 4). By reasoning with
the rule of vector addition, the resulting vector reih is obtained by adding the
expanded vectors r cos h (represented in the horizontal axis) and r sin h
(represented in the vertical axis).
(iii) A mathematical diagram can arguably provide a full-blown proof—at least in
some cases. Consider, for instance, the following theorem from number theory
(Brown 2008, p. 36):
Theorem
1 þ 2 þ 3 þ þ n ¼ n2 =2 þ n=2:
Fig. 4 Mathematical
diagrams can also provide
heuristic tools in proof
construction
Fig. 5 A mathematical
diagrams can arguably
provide a full-blown proof
10 O. Bueno
Proof Admittedly, the diagram only exhibits the result for n = 5. It represents each
natural number as a square, and addition is represented as concatenating and pilling
up such squares. There are 15 squares in the diagram, which corresponds to the
addition of the first 5 natural numbers (n = 5). With a little bit of reasoning about
the diagram, one can see that another way of getting the same configuration of 15
squares would be by squaring n = 5, which would yield a diagram with 25 squares,
and excluding half of them (n2/2), thus yielding a diagram with 12.5 squares;
finally, to complete 15 squares, one adds n/2 (= 2.5) squares back. These opera-
tions, which correspond to the terms in the right-hand side of the identity sign in the
theorem’s statement (namely, n2/2 + n/2), are perfectly general and can be per-
formed for any natural number. Thus, although the diagram itself only exhibits the
result for a particular instance (n = 5), it can be generalized without loss.
It is important to note that the diagrammatic proof incorporates both visual
(pictorial) and linguistic (reasoning-based) traits. While the diagram itself empha-
sizes visual elements, some reasoning about the diagram is needed to establish that
the intended result holds and can be generalized beyond the particular case the
diagram depicts. Diagrams are, thus, hybrid objects that include the visual and the
linguistic, and in the context of mathematical practice, as opposed to the sciences,
they may be used as sources of evidence.
Why do diagrams have such different roles in scientific and mathematical practice?
Diagrams cannot play an evidential role in the sciences given that this requires a
causal relation between the objects in the sample under study and the corresponding
image—such as the one provided by a microscope. It is in virtue of this causal
relation that a micrograph provides evidence for what takes place in the sample.
Given the interaction of the microscope with the sample, the objects in the sample
produce (cause) the microscope image, which, as a result, can be taken as offering
evidence for the presence of the relevant objects. If the image has been properly
produced, it will allow researches to rule out (likely) possibilities that, should they
obtained, the presence of the phenomena displayed in the image would be under-
mined. This process of elimination of undermining possibilities, such as artifacts
and confounding factors, is sometimes achieved by combining the results generated
by the microscope with those of additional instruments.
In the case of the AFM micrograph in Fig. 1 (on the right), researchers
emphasize that it is the presence in the sample of DNA strands configured in a
honeycomb shape that produces the resulting microscope images. Similarly, in the
case of the ‘single’-molecular wire experiment (Fig. 2, on the right), researchers
Visual Reasoning in Science and Mathematics 11
also insist that a current from the surface of the sample to the gold substratum
produces the small blobs on the STM micrographs.2
No such causal relations are required in mathematics. This opens up the pos-
sibility of having diagrams as full-fledged evidential devices, conveying the content
of the theorem’s statement without presupposing any causal relation between the
diagram and the configuration among mathematical objects that the theorem
describes. As opposed to what happens with a micrograph, it is not the case that
relations among mathematical objects produce the corresponding diagram. There is
simply no causal connection among these objects, and none is expected, given that
mathematical objects are causally inert.
Moreover, one need not be a platonist about mathematics to recognize the
potential evidential role of diagrams in mathematical practice. Even if mathematical
objects did not exist at all (as nominalists insist they don’t), diagrams could still be
used as sources of evidence. What is important is that a diagram conveys properly
the relevant conceptual relations described in the statement of the theorem—whether
the objects involved exist or not.3 The non-relevant relations need not be properly
represented at all: diagrams often misrepresent many features of the objects under
consideration. In the case of the number-theoretic diagram just discussed (Fig. 5),
clearly natural numbers are not squares and to add such numbers is not strictly to pile
squares up. Several relations in the diagram cannot, thus, be taken literally. The
diagram is a representational device that does not convey faithfully every aspect of
the relations among the relevant mathematical objects. What is crucial is that the
central, relevant relations—those explicitly stated in the theorem—are properly
displayed in the diagram.
This point also highlights the hybrid nature of diagrams: they are non-linguistic
entities that can have informational content. If the content is right, it may convey all
the information required to establish the truth of the theorem under consideration,
suitably augmented by proper bits of reasoning. By displaying the intended result
(the theorem’s content) in a particular case, and by indicating, by a suitable rea-
soning, the possibility of extending that instance to any relevant case, the diagram
provides the relevant content. This is illustrated in the diagram of Fig. 5 for the case
in which n = 5, since the ability to take the diagram as a source of evidence requires
the additional reasoning that generalizes the result beyond that case.
2
Whether a realist reading of these images is justified or not is not a topic for this paper (see Bueno
2011 for some critical discussion). The point about the evidence requirement stands independently
of this issue.
3
Someone may complain that if mathematical objects don’t exist, then there is no distinction
between proper and improper ways of describing them. Given their nonexistence, it doesn’t matter
how they are described. But this is not right. Sherlock Holmes doesn’t exist, but it is proper to
describe him as a detective rather than a milkman; it is improper to describe him as (literally) a
mouse rather than a person. The same point applies to numbers and other mathematical objects:
even if numbers don’t exist, it is proper to describe the sum of two natural numbers as a natural
number; it is improper to describe the cumulative hierarchy of sets as (literally) a blacksmith.
12 O. Bueno
4
For further discussion of this issue, see Wilson (1995) and Pitt (2004), who quote the passage
from Hooke.
Visual Reasoning in Science and Mathematics 13
The process of learning how to use a microscope requires some care and training
(just as the learning of how to use our eyes does, even if we no longer remember
how that happened). It is by comparing and contrasting a number of views of the
same specimen that one learns how to visualize the object under study.
One issue that Hooke considered was how to present the content of what he
experienced through the microscope. The technology that would allow one to take a
photograph of the visually salient features of the specimen would not have been
created for almost couple of centuries. Hooke had to devise a way of conveying that
information to the readers of his book. He would draw an image of each specimen
he saw after varying the conditions of observation in order to appreciate the con-
trasts and to obtain salient traits of the sample. He would then try to reproduce these
features in a typical image: one that incorporates, into a single image, the variety of
traits of the specimen in order to capture the relevant look of the specimen. The
series of interactions with the sample paves the way to the drawing of a typical case.
The drawing itself is then the evidence that the sample in fact had those features that
the image represents it as having: features that were seen with the microscope at a
certain stage. Thus, in such conditions, so the argument goes, a drawing can be a
source of evidence.
The idea that a mechanically generated image is a source of evidence emerged
from the understanding of objectivity according to which in order to guarantee that
information about the specimen is properly captured and recorded one needs to
implement the process mechanically. This process is expected to ensure that no
illicit interpretation of the data is inadvertently introduced. If, however, there are
mechanisms to guarantee that no unintended interpretation is advanced and that the
data are properly conveyed, then mechanical image generation may not be required.
In this case, a drawing of a typical case can be the source of evidence. However,
these drawings are not diagrams, but are supposed to be faithful representations that
capture the features of the phenomena. They are produced according to specified
rules that ensure that the drawings are sensitive to the visually salient traits of the
phenomenon under study. It is this counterfactual dependence between the speci-
men studied with the instrument and the corresponding drawing that guarantees that
the drawing is a source of evidence. Later, when mechanical image generation
became possible and was established as the norm in scientific research, it is
arguably the counterfactual dependence that supports the resulting images as
evidence (see Bueno 2011).
It may be argued that drawings are used to represent generic types, whereas
machine-made images represent particulars.5 For drawings have a significant trait:
selectivity. Since they are intentionally created, drawings can convey selected
features of their target, and highlight traits that could otherwise be missed. In
contrast, machine-made images simply reproduce those features of the objects that
are present at the scene before the machine and to which the machine is sensitive.
The result is then the representation of the particulars: those before the machine.
5
For a critical discussion of this distinction, see Lopes (2009).
14 O. Bueno
have it, due to their pictorial features, diagrams would tend to convey content about
particulars rather than generic traits.
Some illustrations in archeology, namely, lithic illustrations, also challenge this
divide between drawings and machine-generated images, since these illustrations
are drawings that convey information about particulars (Lopes 2009). Lithic
illustrations are produced in accordance with very strict rules to guarantee that the
relevant features of the stones that need to be studied are properly displayed and the
right traits are highlighted. These rules guarantee the counterfactual dependence
between the traits on the surface of a stone and the corresponding drawing. As a
result, lithic illustrations are taken as sources of evidence in archeology.
Note, however, that lithic illustrations are not diagrams, given that diagrams are
schematic renderings of certain aspects of their targets and need not preserve
visually salient features of the objects they represent. If lithic illustrations were
diagrams, they wouldn’t be sources of evidence, since they need not be produced
counterfactually from the target. In contrast, lithic illustrations are faithful repre-
sentations of their targets, and are sensitive to and properly capture the traits of the
relevant stones.
By considering the relations between the heuristic and the warranting styles
important traits of diagrams can also be highlighted. The warranting style
acknowledges that in order for one to determine that a certain diagram establishes
the intended result, some bit of reasoning is typically required. One needs to reason
with the diagram, in the way illustrated in the discussion of Fig. 5 above, in order to
determine that the relevant theorem does hold.
The warranting style is significant for providing a way of using diagrams in
which they convey evidential information about the objects under consideration.
The heuristic style is important for accommodating the way in which diagrams
provide resources to motivate and understand certain results.
There is a significant connection between these two styles. On the one hand, the
warranting style presupposes the heuristic style in that, in order for diagrams to
provide justification for a given result, they need first to be the kind of thing that can
yield some insight about the subject matter. On the other hand, although the
heuristic style takes diagrams in such a way that they need not provide justification
for the result under consideration, they can still help in the interpretation of certain
results as well as in suggesting and motivating them.
As we saw, central to the divide between the semantic and the received views is the
role of linguistic features in scientific practice. But ultimately both views require
linguistic considerations. Theories, even if thought of as nonlinguistic devices (e.g.
as a family of models), have informational content, which, in turn, involves some
linguistic structure.
16 O. Bueno
In this context, diagrams provide a suitable setting to bring the semantic and the
received views together. Diagrams are non-linguistic devices, but they do have
informational content: (a) they represent a certain situation, (b) and the repre-
sentation is implemented in a given way (according to a certain form). In Fig. 5, the
diagram represents the process of adding n natural numbers as one of pilling up
squares: each square represents a particular number and addition is represented by
the concatenation of piles of squares. In Fig. 2 (on the left), the diagram represents
the sketch of the experiment, with the ‘single’-molecular wire, represented as a
string of bounded atoms, connecting the tip of the STM with the gold basis. In
Fig. 1 (on the left), the expected outcome of the DNA strand experiment is sket-
ched, in which different DNA configurations shaped as triangles, hexagons and
honeycombs are displayed. The particular geometrical objects in the diagram
represent the configuration of the expected DNA arrangements in space.
In all of these cases, the representation is achieved, in part, by a structural
similarity (such as partial isomorphism, partial homomorphism etc.6) between the
diagrams and their targets. It is due to the appropriate structural similarity that the
former can be used to represent the latter in each scientific context. In the case of the
number-theoretic theorem (Fig. 5), the relevant structural similarity is found
between the concatenation of piles of squares and the addition of natural numbers.
In the ‘single’-molecular wire experiment (Fig. 2), there is structural similarity in
the position of the ‘single’-molecular wire relative to the STM tip, the enveloping
substratum, and the gold basis as represented by the configurations on the diagram
and as arranged by the appropriate molecules in the sample. In the DNA strand
experiment (Fig. 1), the structural similarity is established between the geometrical
configurations of lines on the diagrams and DNA strands in the sample.
As noted, diagrams are hybrid entities, very similar—in this respect—to scien-
tific theories. Even if, following the semantic view, such theories are thought of as
families of models (which are nonlinguistic objects) they are models that satisfy
certain conditions, namely, those specified in the formulation of the theories.
Newtonian theory may be presented as a family of models, but these models satisfy
Newtonian laws of motion, which are specified linguistically. Without such lin-
guistic specification it would be unclear what these models are models of. These
conditions need to be presented linguistically, although the particular language in
which they are formulated is, for the most part, immaterial. Similarly, linguistic and
nonlinguistic considerations are brought to bear in thinking through the status of
diagrams. Nonlinguistic considerations enter due to the diagrams’ pictorial nature:
the information they convey is presented in such a way that the shape and relative
position of each line has significance. And it is through their pictorial features that
diagrams have informational content. As we saw above, a sequence of piled-up
squares can be used to represent the addition of numbers, with the number of
squares in the concatenated geometrical object representing the resulting number.
6
For further discussion of these concepts in the context of scientific representation, see Bueno and
French (2011).
Visual Reasoning in Science and Mathematics 17
L’EFFROYABLE RÉVÉLATION
Le temps passa. La tristesse fut à nouveau gravée pour toujours sur les
traits de la fille du bon duc. On ne vit plus désormais ensemble elle et
Conrad. Le duc s’en affligea. Mais avec les semaines successives, les
couleurs revinrent aux joues de Conrad, son ancienne vivacité brilla dans ses
yeux, il continua à administrer le royaume avec une sagesse lucide et
mûrissante chaque jour.
Un bruit étrange, bientôt, se glissa dans le palais. Il grandit, et se
propagea. Les racontars de la cité le répandirent. Il pénétra dans tout le
duché. Et l’on entendait chuchoter: «La dame Constance a donné naissance à
un fils?»
Quand ce bruit parvint aux oreilles du seigneur de Klugenstein, il agita
par trois fois son casque à panache autour de sa tête, en criant:
«Longue vie au duc Conrad! Los! Sa couronne est sûre maintenant.
Detzin s’est acquitté de sa mission. Le brave scélérat a bien mérité sa
récompense!»
Il partit semer la nouvelle au large et au loin. Pendant quarante-huit
heures, il n’y eut pas une âme dans la baronnie qui ne dansât et ne chantât,
ne banquetât et n’illuminât, pour célébrer le grand événement, le tout aux
frais généreux et gais du vieux Klugenstein.
CHAPITRE V
CATASTROPHE ÉPOUVANTABLE