SILFS

SILFS

1
The papers collected in this volume are based on the best contributions to the conference of the Italian Society for Logic and Philosophy of Science (SILFS) that took place in Milan on 8-10 October 2007. The aim of the Society, since its foundation in 1952, has always been that of bringing together scholars working in the broad areas of Logic, Philosophy of Science and History of Science who share an open-minded approach to their disciplines and regard them as essentially requiring continuous confrontation and bridge-building to avoid the danger of over-specialism. In this perspective, logiciansand philosophers of science should not indulge in inventing and cherishing their own internal problems although these may occasionally be an opportunity for conceptual clarication but should primarily look at the challenging conceptual and methodological questions that arise in any genuine attempt to extend our objective knowledge. As Ludovico Geymonat used to put it: [good] philosophy should be sought in the folds of science itself. Contributions are distributed into six sections, ve of which Logic and Computing, Physics and Mathematics, Life Sciences, Economics and Social Sciences, Neuroscience and Philosophy of Mind are devoted to the discussion of cuttingedge problems that arise from current-day scientic research, while the remaining section on General Philosophy of Science is focused on foundational and methodological questions that are common to all areas.

SILFS

New Essays in Logic and Philosophy of Science

SILFS
Editors

New Essays in Logic and Philosophy of Science

Marcello DAgostin Giulio Giorello Federico Laudisa Telmo Pievani Corrado Sinigaglia

Marcello DAgostino Giulio Giorello Federico Laudisa Telmo Pievani Corrado Sinigaglia

Editors

9781848900035_cover.indd 1

28/10/2010 21:07

Table of contents

Editors preface List of contributors PART I LOGIC AND COMPUTING

ix xi 1 3 13 27 37 47

Umberto Rivieccio A bilattice for contextual reasoning Francesca Poggiolesi Reecting the semantic features of S5 at the syntactic level `le Fischer Servi Gise Non monotonic conditionals and the concept I believe only A Carlo Penco, Daniele Porello Sense and Proof Andrea Pedeferri Some reections on plurals and second order logic G. Casini, H. Hosni Default-assumption consequence relations in a preferential setting: reasoning about normality Bianca Boretti, Sara Negri On the nitization of Priorean linear time Riccardo Bruni Proof-theoretic aspects of quasi-inductive denitions Giacomo Calamai Remarks on a proof-theoretic characterization of polynomial space functions PART II PHYSICS AND MATHEMATICS

53 67 81

95 115

Vincenzo Fano, Giovanni Macchia How contemporary cosmology bypasses Kantian prohibition against a science of the universe Giulia Giannini Poincar e and the electromagnetic world picture. For a revaluation of his conventionalism Marco Toscano Besides quantity: the epistemological meaning of Poincar es qualitative analysis

117

131

141

vi

Laura Felline Structural explanation from special relativity to quantum mechanics Miriam Cometto When the structure is not a limit. On continuity through theory-change Gianluca Introzzi Approaches to wave/particle duality: historical analysis and critical remarks Marco Pedicini, Mario Piazza An application of von Neumann algebras to computational complexity Miriam Franchella Phenomenology and intuitionism: the pros and cons of a research program Luca Bellotti A note on the circularity of set-theoretic semantics for set theory Valeria Giardino The use of gures and diagrams in mathematics ` Paola Cantu The role of epistemological models in Veroneses and Bettazzis theory of magnitudes PART III LIFE SCIENCES

153 163

173 183

195 207 217

229 243

Pietro Omodeo Evolution by increasing complexity in the framework of Darwins theory Stefano Giaimo, Giuseppe Testa Gene: an entity in search of concepts Elena Casetta Categories, taxa, and chimeras Flavio DAbramo Final, ecient and complex causes in biology Ludovica Lorusso The concept of race and its justication in biology PART IV ECONOMICS AND SOCIAL SCIENCES

245 257 265 279 289 301

Marco Novarese, Alessandro Lanteri, Cesare Tibaldeschi Learning, generalization, and the perception of information: an experimental study Andrea Pozzali Tacit knowledge and economics: recent ndings and perspectives of research

303

319

vii

Viviana Di Giovinazzo From individual well-being to economic welfare. Tibor Scitovsky explains why (consumers) dissatisfaction leads to a joyless economy Federica Russo Explaining causal modelling. Or, what a causal model ought to explain Enzo Di Nuoscio The epistemological statute of the rationality principle. Comparing Mises and Popper Albertina Oliverio Evolution, cooperation and rationality: some remarks Francesco Di Iorio Self-organization of the mind and methodological individualism in Hayeks thought Simona Morini Can ethics be naturalized? Stefano Vaselli Searles collective intentionality and the invisible hand explanations Sergio Levi The naturalness of religion and the action representation system Andrea Zhok On value judgement and the ethical nature of economic optimality Dario Antiseri Carl Menger, Ludwig von Mises, Friedrich A. von Hayek and Karl Popper: four Viennese in defense of methodological individualism PART V NEUROSCIENCE AND PHILOSOPHY OF MIND

327

347

363 377

389 403 409 423

433

447 463

Fabio Bacchini Jaegwon Kim and the threat of epiphenomenalism of mental states Wolfgang Huemer Philosophy of mind between reduction, elimination and enrichment Laura Sparaci Discourse and action: analyzing the possibility of a structural similarity Alessandro DellAnna Visuomotor representations: Jacob and Jeannerod between enaction and the two visual systems hypothesis

465 481 493

505

viii

Daniela Tagliafico Mirror neurons and the radical view on simulation Vincenzo G. Fiore Multiple realizations of the mental states: hunting for plausible chimeras Arturo Carsetti The embodied meaning and the unfolding of the minds eyes Katja Crone Consciousness and the problem of dierent viewpoints Giulia Piredda The whys and hows of extended mind Carmela Morabito Movement in the philosophy of mind: traces of the motor model of mind in the history of science Jean-Luc Petit The brain, the person and the world PART VI GENERAL PHILOSOPHY OF SCIENCE

515 529 539 547 559

571 585 601

Gustavo Cevolani, Vincenzo Crupi, Roberto Festa The whole truth about Linda: probability, verisimilitude, and a paradox of conjunction Antonino Freno Probabilistic Graphical Models and logic of scientic discovery Massimiliano Carrara, Davide Fassio Perfected science and the knowability paradox Luca Tambolo Two problems for normative naturalism Silvano Zipoli Caiani Explaining the scientic success. A critique of an abductive defence of scientic realism Mario Alai Van Fraassen, observability and belief Marco Giunti Reduction in dynamical systems: a representational view Alexander Afriat Duhem, Quine and the other dogma Edoardo Datteri Bionic simulation of biological systems: a methodological analysis Luca Guzzardi Some remarks on a heuristic point of view about the role of experiment in the physical sciences

603 617 629 637

649 663 677 695

711

725

Some remarks on a heuristic point of view about the role of experiment in the physical sciences
Luca Guzzardi

Introduction

One major contribution of the 20th century epistemology is the focus on a non-na ve concept of experiment. We have learnt experiment does not provide either a verication or an immediate falsication of a theory; physical observations are theory-laden and the experimental praxis follows the theory and is at the expenses of a silent, long and slow theoretical work. Otherwise the experimental praxis wouldnt simply exist, and the work of experimenters is by far and in the rst place a theoretical one. We can trace this attitude back to Pierre Duhems Th eorie physique (1906) and Karl R. Poppers Logik der Forschung (1934; Logic of Scientic Discovery, 1959). Associated with this point of view, which seems in itself very dicult to question, one can often nd the unexpressed assumption that experiments are control means of theory, so they have no other purpose than to test and control the correctness of a system of empirical statements. According to this, any other role experimentation could play has so little importance in the eyes of epistemologists that it is at best a collection of interesting peculiarities (and a waste of time in the worst case). This point of view is appropriately summarized in the incipit of Poppers Logic of Scientic Discovery :
A scientist, whether theorist or experimenter, puts forward statements, or systems of statements, and tests them step by step. In the eld of the empirical sciences, more particularly, he constructs hypotheses, or systems of theories, and tests them against experience by observation and experiment [15, p. 3], [16, p. 27].1

In this paper I will try to show how this view, that emphasizes the prominence and prevalence of theory against experiment, leads to overlook a major use of experiments, provided with strong historical evidence and supported by epistemological arguments. In the picture I will try to give, experiments do not play the main role of empirical control of theory, though they can be used with that function. But I shall argue that their main role is a dierent one.
1 Note that the words whether theorist or experimenter were added in the English edition 19581959 (and were not reported in the German edition 1966), as if Popper felt a need of specication or emphasis: the experimenter must do the same work, follow the same procedure and more importantly apply the same methods as the theorist colleague

726

Luca Guzzardi

A cautionary argument is to be made in advance: I do not claim my suggestions will apply to any eld of empirical sciences, though I think this would became arguable, provided that appropriate changes in my picture are introduced. However, I dont try here to embark on this enterprise. This is also the reason why historical examples in this paper are only taken from the history of physics and my suggestions are restricted to the physical sciences.

Proof: a legal concept and its epistemological pendant

In which sense does an experiment become an experimental test ? How did this idea that a certain experience is actually a proof against or in favour of a certain belief, conviction or system of hypotheses acquired relevance till to the point it has become obvious? An answer to this could be nd in a brief account of the history of the concept of proof. The term proof traces most probably back to legal jargon. Generally speaking, to describe something as a proof means to establish a criterion for deciding about something: proof is what provides the basis of our decision and justify it. So, for instance, the verdict of not proven, typical of the Scots Law, means an acquittal for insucient proof, because evidence is so inadequate and defective that a judge would have no reason to convict a defendant. This legal background, which transpires from the humean (and than kantian) image of a court of justice for the reason,2 involves sometimes pure scientic work as well: this is the case of William Thomsons (Lord Kelvin) assessment of a particular and very sosticated theory of aether he himself contributed to developed some years before: I am thus driven to admit, in conclusion, that the most favourable verdict I can ask for the propagation of laminar waves through a turbulently moving inviscid liquid [i.e. the ether] is the Scottish verdict of not proven [19, p. 352]. According to Kelvin, the problem with this theory was not the rened mathematics developed by him and others, but merely the experience, which didnt suce to support the theory. The legal imprint of the term proof still sticks on this concept even if we move away from legal jargon.3 In an arithmetical textbook dated around 1430
2 See [11, p. 3], [12, A11, eng. tr. p. 101]. Not to mention what Kant argues about the legal sense of the concept of (transcendental) deduction in 13 of the Critique of Pure Reason. A good understanding of the concept of proof, its development and background also in legal terms is provided by [6] (see more in particular about Kant and Hume [6, pp. 1523 it. tr.]; about proof and the legal tradition [6, pp. 3543 it. tr.]). 3 Very interestingly the concept of proof seems originally have been aected by religious tradition too. One of the rst occurrences of the term proof (intertwined with its legal meaning) you can nd in English is contained in a Middle-age text known as Ancren Riwle or Ancrene Wisse, a Regula for Anchoresses written around 1200 by an anonymous English churchman for the instruction of a small community of three women about to become religious recluses. In this book we can nd one of the rst occurrences of the word in English. What in the eyes of the anonymous author justies rules (therefore providing proofs) are mostly stories from the Bible. So in a couple of passage the speaker who gives the rule to the three young ladies argues: That this is true [...] here is the proof [preoue ] [1, pp. 52, 53] and further: Because I said that we nd this both in the Old Testament and also in the New, I will, out of both, show an example and proof [1, pp. 154, 155]. The term proof is used here in the sense of what makes good that is proves a statement:

Role of experiment in the physical sciences

727

and called The art of nombryng (a translation of a Latin textbook De arte numerandi, written in the 13th century and attributed to John of Holywood), proof [proue ] takes explicitely the meaning of a test or a trial to check the correctness of an arithmetical calculation: The subtraccioun is none other but a proue of the addicioun, and the contrarye in like wise[2, p. 6]. In fact, this sentence, despite the triviality by which it is only seemingly aected, brings to bear what I called the legal origin of the concept of proof. Inverse operations in Mathematics are treated as proofs because they allow to decide to judge about the correctness of calculi, providing a justication of them. Nevertheless, by this way the arithmetical proof achieves somehow a new feature, namely the feature of a trial a term involving an obvious legal background. Like in a lawsuit, the fact to have overcome a trial (the subtraction proof for the addiction, for example) puts de iure the calculation into the realm of what is legitimate, so to speak, beyond reasonable doubt. As in Humes and Kants metaphor of the court of justice for human reason, on this theoretical level proof can indicate both the individual evidences used to convince the mind and the entire process of convincing someone a process sometimes called demonstration, that is the ability to deducing something from certain, denite assumptions produced before the attentive and watchful eye of understanding. The relationship between proof and demonstration was developed by John Locke (in Book IV of his Essay Concerning Human Understanding, 1690). He put emphasis on the rst one, with an important reference to the process of vision:
Those intervening Ideas, which serve to shew the Agreement of any two others, are called Proofs ; and where the Agreement or Disagreement is by this means plainly and clearly perceived, it is called Demonstration, it being shewn to the Understanding, and the Mind made see that it is so [13, Book IV/ii, 3, p. 532].

A proof builds up a demonstration in so far as it brings to evidence what, as beeing seen, takes its own place within positive and true knowledge. Something that has achieved this status of a demonstrated and established truth can never be moved away from that place by any kind of contrary judgement. Indeed, the understanding see it and the Mind made see that it is so. But Locke points out all this has no relation with experience, both crude or controlled in the form of experiment. According to him, the way of nding thruths by proofs and demonstrations applies only to ideas and even what he calls the Art of nding Proofs is one major breakthrough of an argumentation style which is to be learned in the Schools of Mathematicians, who from very plain and easy beginnings, by gentle degrees, and a continued Chain of Reasonings, proceed to the discovery and demonstration of Truths [13, Book IV/xii, 7, p. 643]. On the contrary, experiments would absolutely be not able to yield established knowledge. Though they have a dierent and very important role:
A Man accustomed to rational and regular Experiments shall be able to see farther into the Nature of Bodies, and guess righter at their yet unknown Properties, then one, that is a stranger to them: But Yet, as I have said, this is but Judgement and Opinion, not Knowledge and Certainty [13, Book IV/xii, 10, p. 645; italics mine]. an evidence that is sucient or contributes to establish anything; for instance, a rule for living.

728

Luca Guzzardi

A patient experimental culture

Locke remarkably reduced the signicance of experiment in the sense of a proof: because of the weakness of achieving knowledge by means of experience, every test we undertake even to falsify (not to say to verify) our theories is not but judgement and opinion. Above all, what is important in Lockes idea about experimenting nature is not that experiments are means to prove or test our knowledge, but that they are a substantial part of a heuristics, so they can lead and ultimately help grow our knowledge. Of course, we need to examine step by step any individual case of a phenomenon to be sure that our principle will carry us quite through, and not be as inconsistent with one Phoenomenon of Nature[13, Book IV/xii, 13, p. 648]. However, this is nothing but a side eect of a preliminary be accustomed to rational and regular Experiments, that lead our opinion and compel us to some guesses. Therefore, experimentation is the source of the whole process of guessing. This aspect reaches far beyond Lockes empiricist perspective because it doesnt involve merely the logic of scientic discovery; it rather deals with both the praxis and the practice of science: such a specic praxis as we can by and large outline it. As well known, Locke had close relationship with the environment of the Royal Society.4 According to its Statute frequent meetings of the fellows if possible once a week should be hold. The business of their weekly Meetings shall be, to order, take account, consider, and discourse of Philosophical Experiments, and Observation [18, p. 145]5 where the expression philosophical experiments means nothing but experiments in natural philosophy, i.e. scientic experiments as we are used to call them. Remarkably, to make experiments an activity that Bishop Thomas Sprat (who was himself a fellow and one of the rst historians of the Royal Society) regarded as the substantial part of the meetings was not the business of any Fellow. The Constitution of the Society stated that a person should be selected for this purpose and his exclusive duty had to be providing for experiments. The name of such employee was the Curator of experiments. While the Fellows came from the most dierent occupations and cultivated natural philosophy as their personal interest and passion, however serious these might be, curators were professional experimenters and they got by the Society i.e. by the Fellows themselves a salary for their work (the Statute established a maximum of two hundred pounds per year, but the rst curator, Robert Hooke, got only thirty pounds in addition to an apartment in the buildings of the Royal Society).
was elected a fellow of the Royal Society in November 1668 (six years later its foundation). It is known he was quickly appointed to a committee for experiments and twice served on the council, but apparently he has little contributed to the work of the Society. Nevertheless, he dedicated the Essay, rst published in 1690, to his friend Thomas Herbert, then President of the Society; throughout his life, either in England or in exile, Locke followed with interest the Societys activities. About the relations between Locke and the Royal Society with regard to the purposes of present paper see [3, vol. 1, pp. 245249], [17, pp. 7374], [21, pp. 5258], [23, pp. 3638], and [22, pp. 1723]. About Lockes experimental Knowledge see [4, 204209]. 5 Sprats quoted text follows the edition London 1667.
4 Locke

Role of experiment in the physical sciences

729

The oce of a curator must have been a delicate one, as the complicated appointment procedure points to: to begin with, unless an eminent person was known to the Fellows for his competence and worth, usually curators had to be examined very carefully by the Society before election. They were rst recruited for a trial period, which normally didnt take a full year, at the end of which they shall be either elected for perpetuity, or for a longer time of probation, or wholly rejected [18, p.147]. Specically, their business was to
take care of the managing of all Experiments, and Observations appointed by the Society, or Council, and report the same, and perform such other tasks, as the Society, or Council shall appoint: such as the examining of Sciences, Arts, and Inventions now in use, and the bringing in Histories of Natural and Articial things, & c.[18, p.147].

The Statute provided also a brief sketch of a typical curator: his background and competence had to be appropriate to the oce, he had to be skilled in Philosophical, and Mathematical Learning, well versed in Observations, Inquiries, and Experiment of Nature and Art [18, p.147]. In other words, curators were essentially technicians, who had of course to know the elements of natural philosophy and mathematics and, so to speak, feeling at home in applying them. However, this knowledge was rst of all intended for their job, i.e. managing experiments before an audience consisting of the Fellows of the Royal Society. According to Locke, the most valuable gift of a good philosopher trained at the School of Mathematicians, who proceed by proofs and demonstrations (and maybe refutations) of their own ideas is perhaps sagacity, that is the ability to nd out quickly and without delay a basis for the arguments he want to use, providing them with an indubitable certainty. The most important quality of a curator of experiments must have been, instead, the innite patience he needed to vary every time his experiments and carefully examine a profusion of individual cases. After all, any experimenter knows that everything is nothing but an individual, unique case... Natur ist nur einmal da , as Ernst Mach put it in The Science of Mechanics : nature is but once there. But after all varying experimental conditions of a phenomenon means to show, ad excludendum, what does not change in variation and how things generally are, so that one can illustrate through experiments the typical behaviour of things. Therefore, curators were required to have a great deal of imagination, both for observing nature in so many aspects as possible and for illustrative, teaching purposes. The aim of experiments in public demonstrations (at the Royal Society, for example) was not only to show nature, but also to illustrate theories in so many ways as possible. Like in Lockes account, for learning science is needed to see and discuss experiments mainly through the ability of a good experimenter. Remember that according to Locke experiments had a very important role in teaching: people accustomed to rational and regular Experiments shall be able to see farther into the Nature of Bodies, and guess, and so on. Experiments acted somehow as a propaedeutics to research, for young and less young natural philosophers (like the Fellow of the Royal Society) could learn the art of

730

Luca Guzzardi

seeing farther into the Nature of Bodies and make guesses about their properties. But no illusions about that: in Lockes eyes experiments provide very restricted control, in most cases they say nor yes neither no and there is no assurance of an unquestionable verdict from them. So experiments would play their major role in giving so to speak a sensible illustration of theories and training scientists about them. Often during 17th and 18th century experimenters were called demonstrators in scientic elds, meaning with that laboratory professionals who gave public demonstrations of an experiment for teaching purposes before a more or less large audience. Their aim was not the same of the theorists (or something like that the theorists could expect by experimenters), i.e. to test a set of hypotheses. They rather aimed to provide what I called a sensible illustration of a theory, so that people could learn and discuss that theory. In order to perform this, the experimenter-demonstrator, an institutional version of which is embodied by the curator of experiments at the Royal Society, did not have to take critical attitude against theories he was demonstrating. He had to be a loyal supporter and a strenuous defender of these; he had to be convinced of these in order to convince his audience. In this regard, by no means an experimental demonstration, either in Poppers or in Lockes sense, would provide a proof for or against a theory. But a demonstration illustrates a theory6 and in the meantime it can also give an apology, so the experimenter properly plays the role of an apologist. One may think, I actually am trying here to make an apology of something namely of the inductive method, that is inferring theories from experiments. But to state an apology through experiments in the sense I have just described does not involve an application of inductivism. Experimenters as I have described them scientists of a well recognizable kind in the history of science dont bother stating a (new) general theory starting from empirical data. Scientists as Robert Hooke, Francis Hauksbee or Jean-Th eophile Desaguliers all of them Curators of experiments at the Royal Society and many others, when they act as apologists (and uniquely in this case), only take care of defending theory against possible assaults by enemies and indels and preventing potential heretics from proselytizing.7 To state an apology is also very dierent from performing an experimental test of a theory: in last case, even if we are expecting a positive result in favour of a certain theory, in principle an error could occur that is the fallibilistic point of view. But apology does imply that there is no doubt that theory is fundamentally correct (though some details could vary).8 Therefore, there is no need of any control at all. And this applies in principle : the possibility something might be wrong in the very fundament of the theory at issue is not even taken into account. Apology implies two actions are basisense of scientic demonstration was probably relevant for Kant himself. On this issue see [14, pp. 9597]. 7 Sometimes they are driven in doing this job by theorists or eminent scientists, as it happened in the case of Newton and his curator Francis Hauksbee. See [10, pp. 229234]. 8 I dont need to point out that often experiments can be designed in order to extend a given theory. As well known, this feature has been emphasized by Bas van Fraassen, whose point of view I discuss in the third section of present paper.
6 This

Role of experiment in the physical sciences

731

cally needed: defending and convincing. A theory will not be conrmed nor corroborated in Poppers sense, which implies a reference to a (fallibilistic) degree of condence that doesnt occur here. A theory will rather be literally conrmata in Latin meaning of the word: experiments conrm a theory in the sense that they make it rmer, more solid against possible assaults than it would be without them. No matter of corroboration degree; no matter of psychology or convictions. This is something objective. As historians of science have recognized since some decades, the rapid success of Newtonianism and his general acceptance as the standard view during 18th century throughout the Continent was also due to an ecient apologetics. Amongst its major advocates are to be mentioned Dutch authors as Willem Jacob sGravesande and Pieter van Musschenbroek. Following Hauksbees and Desaguliers experimental tradition, sGravesande wrote in 1721 a physics textbook addressed to students that very quickly became (with its third edition, 1724) one of the most important and inuent treatises of his age. Its Latin title remarkably was Physices Elementa Mathematica, experimentis conrmata sive Introductio ad philosophiam Newtonianam i.e. The Mathematical Elements of Physics Conrmed by Experiments. What is really impressive about this book is the profusion of experiments compared with the small number of pages devoted to mathematical scholii. Its also not surprisingly that the rst English translation of this textbook (1725) was made by a curator of the Royal Society, namely Jean-Th eophile Desaguliers.9

Two cultures of experiment and a heuristic point of view on its role

I have contrasted the usual concept of experiment as a control, which I described using analogies from a legal background (but note how many and powerful analogies with originally legal concepts Popper uses in Chapter 5 of The Logic of Scientic Discovery, namely The Problem of the Empirical Basis),10 with a broader concept, where experiments have at least a threefold role: rst, they are illustrations of a theory and by this way can provide a good deal of examples for (in second place) teaching and (in third place) defending the theory itself. These dierent nuances about the role of experiment are of course related and interacting issues. They share at least one thing: up to this point experiments tightly remain in the hands of theorists, who need to illustrate and defend their creatures and perhaps to nd and teach followers. And of course the theorist (not the experimentalist, pace Popper) feels the
9 For further details see [8]. More in particular about sGravesandes and J.-T. Desaguliers role in defending and conrming Newtonianism, see [7, pp. 9697]. 10 So for example: The verdict of the jury (vere dictum = spoken truly), like that of the experimenter, is an answer to a question of fact (quid facti ?) wich must be put to the jury in the sharpest, the most denite form. Than, Popper makes clear that what question is asked, and how it is put, will depend very largely on the legal situation, i.e. on the prevailing system of criminal law (corresponding to a system of theories). And again: In contrast to the verdict of the jury, the judgement of the judge is reasoned; it needs, and contains, a justication. The judge tries to justify it by, or deduce it logically from, other stetements: the statements of the legal system, combined with the verdict that plays the role of initial conditions [16, pp. 109110].

732

Luca Guzzardi

need to test his theory and exploits the ability of skilled experimenters for this purpose. In addition to this, following Locke I have argued that experiments can provide what we could properly call a heuristics. Though this last feature might be related with the three nuances I have mentioned above, I want to suggest that providing a heuristics means something very dierent. I will try to explain what the dierence is. At the origin of modern physics there might be a distinction both inuent and elusive between two dierent kinds of experimental culture, which also shaped dierent research styles and dierent ways to consider science and scientic methods. One culture thinks of experimentation as if it would only be depending on theories and points out that it provides a control for theories. Following a suggestion by Ian Hacking, I shall call this the theoretical approach to the experiment.11 Supporters of this culture mostly ignore other features of experimentation, which on the contrary are as crucial as the only one aspect they emphasize, though maybe less elevated (such as the illustrative-teachingdefending role of experiment). Therefore, they can regard an experiment as a proof in favour of or against a theory. Maybe this rst approach to experiments, with the concept of proof as its legal pendant, arose in courtrooms, thanks to the job of brilliant orators, lawyers and judges; then it settled from the dusty reading stands of the universities on the aseptic writing desks of natural philosophers and nally reached the chairs of the fellows of titled institutions like The Royal Society. Roughly in the meantime another culture of experiments, which I shall call the experimental tradition, was growing both in craftsman laboratories and in the rst machine shops, amid dusty workbenches, scraps of unsuccessful experiments and instruments without any apparent utility, unawarely developed by scarcely educated people who have no fear to dirty their own hands. Of course we can nd this tradition in the same rooms of the Royal Society; but their representatives did not sit in that educated and mixed audience. The experimental tradition rather passed through the skilled hands of clever professionals like the Curators of experiments: people with technical-practical background who had to perform experiments before that audience. For them experiment was synonymous with uncertainty and doubt, because we only faced with individual cases, and no proofs can be made from experiments. But in their hands experiments, if treated with care, could provide a useful heuristics, so that the experimenters could make (and let make to their audience) some guesses about the nature of bodies, to put it again in Lockes terms. An attentive intervention by a skilled and patient experimenter could and did open new ways to research, driving his guesses and those of his audience. To do this one needs a very peculiar kind of knowledge. I argue that the design and implementation of an experiment requires more craftsmanship than pure, theoretical and objective knowledge: it involves an actual manipulation
11 Hacking refers to experimental and rational faculties complaining Popper and Lakatos, amongst others, because they emphasize only the rational faculty (that is the style I term here the theoretical approach to the experiment) [9, pp. 260261].

Role of experiment in the physical sciences

733

of the world, based not only on knowing something, but most of all on knowing how things are to be made and, last but not least, on personal experience. In other words experiments are a pragmatic matter, indeed matter of praxis and practice. In The Aim and Structure of Physical Theory Pierre Duhem gave a colourful and vivid idea of this sort of knowledge, when he claimed what a researcher has to know to enter a laboratory. Duhems suggestion is, maybe consistently with the authors will, often read as an example of theoryladeness of observation or experiment. But the theory of which experiments would be laden with is of this very peculiar kind: it is, so to speak, a theory-intended-for-the-experiment. The theoretical work of an experimenter, if there is something like that, does not simply deal with abstract notions. This kind of knowledge is meaningful only in the context of an actual making experiments. It is a practical, applied knowledge of how instruments are to be used, how devices are to be manufactured, how they as experimenters have to dirty their own hands. This pragmatic knowledge this knowledge of a theory-intended-for-theexperiment plays an extremely important role in the actual implementation of experiments. I have emphasized that one of the most important qualities of an experimenter is patience in varying experimental conditions to show what does not change and what is to be considered the typical behaviour of things, that is the way things generally are. That was for exemple the case of Francis Hauksbee, the curator of experiments of the Royal Society whose appointment in 1703 was heavily supported by Newton himself as the President of the Society: his ability in varying experimental conditions became legendary. In particular, it was this ability that led him to his remarkable results in examining some subtle electrical phenomena as the luminosity of phosphorus in a Torricelli vacuum.12 From this point of view making experiments is very dierent from simply testing (controlling) a theory, although experiments can be used for such a purpose. And yet, this doesnt suce to claim for a heuristic role of experiments. According to a well-known witty remark ` a la Clausewitz, dued to Bas van Fraassen, experimentation is nothing but continuation of theory construction by other means. However, van Fraassen is not willing to admit that experiments are [...] designed to test theories, to see if they should be admitted to the oce of truth-bearers [20, p. 73].13 Experiments, he points out, rather aims to ll in the blanks in a developing theory, so a whole theory construction can make advance. In fact, as Ian Hacking has observed, this lling in the blanks becomes the major issue of experimentation in van Fraassens account of the scientic image: indeed, lling in the blanks [means] guiding the continuation of the construction, or the completion, of the thoery. On the other hand, the theories formulate the questions worth being answered and embody a guiding factor in the design of the experiments to answer those questions [20, p. 74].14 Experiments are always theory-dependent and theories come rst. By no means an experi12 About 13 About

Hauksbees experiments see the quoted work by [10, p. 230]. his famous Clausewitz-styled statement see [20, p. 77]. 14 For a criticism see also [9, pp. 238240].

734

Luca Guzzardi

menter could suggest to a theorist a new insight through his own work: in van Fraassens account, even if experiments are not only tests for empirical adequacy of the theory as developed so far, nevertheless they cannot be anything but useful blank-llers for a theory could make progress. According to van Fraassen, this being-for-a-theory of any experiment (as opposed to the theory-intended-for-the-experiment outlined above) should entirely full and exhaust the role of the experiment. Though experiments could help to construct or complete a theory, its up to the theory to provide a heuristics. However, this seems not to be consistent with historical evidences. Not only the case studies van Fraassen refers to are prone to a quite dierent interpretation, as shown by Hacking. In addition, his view can neither explain the relevance of professional experimenters as the Curators and the function the Statute of the Royal Society stated for them, nor account for the threefold role of experiment (illustrating-teaching-defending theories) I discussed above. Finally, van Fraassens perspective doesnt take into adequate account that sometimes experimentation can be the guiding factor which come rst in the design of a specic theory. The double-slit experiment provides a good exemple of that. In XIX century this was supposed to be (in fact, it was) a crucial experiment between corpuscular and wave theory of light; but in XX century the double-slit experiment became a major issue for quantum mechanics; and here his role as a test is evanescent against his role as a heuristic tool. Look at the use Richard Feynman makes of it in his presentation of quantum mechanics: it is a heuristic one a heuristics which indicates the most important features of quantum theory. Note that it doesnt matter here if this experiment is actually a Gedankenexperiment, but only the role it plays. As Feynman states: We are doing a thought experiment, which we have chosen because it is easy to think about. We know the results that have been done, in which the scale and the properties have been chosen to show the ects we shall describe [5, vol. 3, pp. 16]. Just in the same way a skilled experimenter as Francis Hauksbee chose, so to speak, the scale and the properties of phenomena to show the ects he aimed to describe: in other words he chose a heuristics. This is not lling in the blanks of a guiding theory behind the experiment; this is simply a good example of a theory-intended-for-the-experiment (or of an experimental faculty, to put it in Hackings terms). According to Feynman, the double-slit experiment has in it the heart of quantum mechanics. In reality, it contains the only mystery. We cannot make the mystery go away by explaining how it works. We will just tell you how it works. In telling you how it works we will have told you about the basic peculiarities of all quantum mechanics[5, vol. 3, pp. 113]. So the experiment is a little bit more than a blank-ller in the theory constuction of quantum mechanics: in reality, it embraces the whole theory. Feynman even refuses any theoretical explanation behind the phenomenon: One might still like to ask: How does it work? What is the machinery behind the law? No one has found any machinery behind the law. No one can explain any more than we have just explained. No one will give you any deeper representation

Role of experiment in the physical sciences

735

of the situation. We have no ideas about a more basic mechanism from which these results can be deduced [5, vol. 3, pp. 113]. Explanation here simply raties in the unchangeable statement of the law that typical, paradigmatic behaviour of things the experiment has found, shown and demonstrated. To put it in Lockes terms, the double-slit experiment helps guess righter at [...] yet unknown properties of nature and Feynman still uses it to shape his path integrals formulation of quantum mechanics. So, shall we maybe reverse van Fraassens claim? Is really experimentation the continuation of theory construction by other means? Or shouldnt we admit, without burding this with any inductivist nuance, that theory is often the prosecution of experiments with other means (and experiments sometimes provide a rst glimpse of a theory)? Maybe Popper is right to argue, quoting Novalis, one has to cast nets to catch sh and hypotheses are our nets. We need audacious, open-minded hypotheses, marvellously imaginative and bold conjectures, to put it in Poppers terms. I would like to suggest that one way to make such marvellously imaginative and bold conjectures is indeed to make experiments.

BIBLIOGRAPHY
[1] The Ancren Riwle. A Treatise on the Rules and Duties of Monastic Life. Edited and translated from a semi-saxon manuscript of the 13th century by J. Morton, printed for the Camden Society, London 1853. [2] The art of Nombryng. A translation of John Holywoods De arte numerandi. In The Earliest Arithmetics in English. Published for the Early English Text Society (Extra Serie, n. 118), Oxford University Press, Oxford 1922. [3] H.R.F. Bourne. The Life of John Locke. Scientia Verlag, Aalen 1969 (Reprint of the Edition London 1876). [4] J. Colman. Lockes Theorie der empirischen Erkenntnis. In John Locke, Essay u ber den menschlichen Verstand. Herausgegeben von U. Thiel, pages 197221. Akademie Verlag, Berlin 1997. [5] The Feynman Lectures on Physics. Addison-Wesley, Reading 1964. [6] F. Gil. Provas. INCM, Lisboa 1986. It. tr. Prove. Attraverso la nozione di prova/dimostrazione. Jaca Book, Milano 1990. [7] G. Gori. La fondazione dellesperienza in sGravesande. La Nuova Italia, Firenze 1972. [8] W.J. sGravesande. Physices Elementa Mathematica, experimentis conrmata sive Introductio ad philosophiam Newtonianam. Leidae 17201721. [9] I. Hacking. Representing and Intervening. Introductory Topics in the Philosophy of Natural Science. Cambridge University Press, Cambridge 1983. [10] J.L. Heilbron. Electricity in the 17th and 18th Centuries. A Study of Early Modern Physics. University of California Press, Berkeley 1979. [11] D. Hume. A Treatise of Human Nature. Edited by D. Fate Norton and M.J. Norton. Oxford University Press, Oxford/New York 2000. [12] I. Kant. Kritik der reinen Vernunft [1st edition]. In Kants Gesammelte Schriften, herausgegeben von der K oniglich Preussischen Akademie der Wissenschaften, erste Abt., Bd. IV. Reimer, Berlin 1911. Eng. tr. Critique of Pure Reason [The Cambridge Edition of the Works of Immanuel Kant]. Cambridge University Press, Cambridge 1998. [13] J. Locke. An Essay Concerning Human Understanding. Edited with an introduction, critical apparatus and glossary by P.H. Nidditch. Clarendon Press, Oxford 1975. [14] F. Moiso. Morfologia e losoa. In Annuario Filosoco. Mursia, Milano 1992. [15] K.R. Popper. Die Logik der Forschung. J.C.B. Mohr (Paul Siebeck), T ubingen 1934, 1966. [16] K.R. Popper. The Logic of Scientic Discovery. Hutchinson & Co., London et a. 1959, 1968. [17] M. Purver. The Royal Society: Concept and Creation. Routledge and Kegan Paul, London 1967.

736

Luca Guzzardi

[18] T. Sprat, The History of Royal Society of London For the Improving of Natural Knowldege, edited with critical apparatus by J.I. Cope and H.W. Jones, Routledge and Kegan Paul, London 1959. [19] W. Thomson. On the propagation of laminar motion through a turbulently moving inviscid liquid. In The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science, 24, 1887. [20] B. van Fraassen. The Scientic Image. Clarendon Press, Oxford 1980. [21] J.W. Yolton. Locke and the Compass of Human Understanding. A Selective Commentary on the Essay. Cambridge University Press, Cambridge 1970. [22] P. Walmsley. Lockes Essay and the Rethoric of Science. Bucknell University Press, Lewisburg and Associated University Presses, London 2003. [23] R.S. Woolhouse. Locke. The Harvester Press, Brighton 1983.