1. Introduction

Michael Faraday's independent research in electromagnetism began when his friend Richard Phillips asked him to write an account of the origins and developments within electromagnetism over its roughly 2-year history as a recognized scientific domain (see letters 145, 147, 157, and 200 in James Reference James1991). In order to write an adequately informed history, Faraday replicated Hans Christian Oërsted's experiments leading to the discovery that voltaic electricity produces magnetic deflections and the experiments occasioned by this novel experimental finding. The resulting work—the “Historical Sketch of Electro-Magnetism” (hereafter, “Historical Sketch”)—was published anonymously in the Annals of Philosophy (Faraday Reference Faraday1821–22). Perhaps the best characterization of this work is that it was an extensive review of the state of electromagnetism, including descriptions of the (i) significant phenomena, (ii) pressing research questions, (iii) experimental apparatus, (iv) theoretical developments, and (v) leading figures contributing to this domain.

The “Historical Sketch” is a fruitful source for exploring the epistemic significance of experimentation. Faraday's sustained criticisms of theoretical speculation within electromagnetism demonstrate significant diffidence concerning proposed accounts of the underlying productive causes of phenomena. But the “Historical Sketch” also evidences a nuanced understanding of the dynamics between experiment and theory, especially at the earliest phases of research within a domain. Additionally, Faraday describes methods that enable experimentalists to dissociate experimental results from the theoretical commitments generating their research.Footnote ¹ In this article, I argue that these dissociative methods enable experimentalists to avoid a form of theory-dependence I call framework theory-dependence.Footnote ² An experiment is framework theory-dependent when its processes and the descriptions of experimental results derive their meaning from the particular theoretical framework generating the experimental inquiry. In Section 2, I discuss Faraday's distinct uses of the term ‘theory’ in the “Historical Sketch” and his theoretical diffidence concerning theoretical views. In Section 3, I classify the variety of experimental processes Faraday catalogs, distinguishing broadly between theory-guided and exploratory experimentation.Footnote ³ In Section 4, I develop and defend Faraday's account of the dissociative methods by which one can detach successful experimental results from framework theory-dependence. In Section 5, I consider whether these dissociative methods are effective in confronting other forms of theory-dependence.

2. Faraday's “Historical Sketch” and Some Varieties of Theoretical Explanation

Faraday's primary goal in writing the “Historical Sketch” was to provide a framework for understanding the various experimental discoveries in the early history of electromagnetism. The first two sections summarized Oërsted's experimental findings and the various discoveries of fact produced by the experiments building on Oërsted's work. Faraday devoted the third section to a systematic critique of the theoretical views proposed to explain this rapidly expanding set of electromagnetic phenomena. This structure was intentional and reflects the fact that Faraday had “found much difficulty in gaining a clear idea of what had been done, and by whom, in consequence of [the] great variety [of published experimental findings], the number of theories advanced in them, their confused dates, and other circumstances” (Reference Faraday1821–22, 195). So, he attempted to clarify what had been discovered independently of the theoretical views that had grown up alongside the experimental discoveries.

But describing the facts as they were discovered required Faraday to acknowledge the vital role of theoretical speculation in scientific inquiry. For instance, he focused much of his discussion on a careful articulation of the discoveries of fact generated by Ampère's theoretical claim that magnetism is caused by the activity of electrical currents. As such, Faraday clearly recognized that theory motivates and guides experimental research and that experimental testing can confirm theoretical predictions concerning novel phenomena and apparatus. Nonetheless, he worried about premature theoretical speculation and the potentially misleading implications of describing experimental results with theoretically infused terminology.Footnote ⁴

Although he does not provide an analysis of theoretical terminology, his systematic critique of existing theoretical conjectures suggests a distinction between the following kinds of theoretical claim:

1. 1. Vague metaphysical commitments concerning the interconnections between various domains of the sciences.

2. 2. Hypotheses concerning the underlying states of various elements of an experimental system.

3. 3. Statements of the laws governing these hypothetical underlying states.

4. 4. Theoretical claims concerning the agencies actively producing phenomena, including proposals concerning the processes by which these agencies produce phenomena and the underlying material nature of these agencies.

Several observations should be made about these distinct kinds of theory. First, vague metaphysical commitments are valuable primarily because they motivate and direct experimental research programs. Referring to Oërsted's work, Faraday (Reference Faraday1821–22, 107) writes, “It cannot be doubted for a moment by any one who has read the papers of this philosopher both on the discovery and prior to it, that his theory rather led to the experiments, than the experiments to the theory.” Experimental results function to refine and clarify the nature of these metaphysical commitments. Second, hypotheses concerning the underlying states of elements within an experimental system are not necessarily dependent on theoretical views concerning the underlying agencies producing these hypothetical states. And experimentalists often propose these hypotheses merely as an account of a particular phenomenon under investigation rather than a whole range of phenomena. Third, proposed laws governing hypothetical states of elements within an experimental system constitute a theoretical advance beyond the supposition of these underlying states. But they are also distinct from any theory concerning the underlying agencies responsible for producing the phenomena in question. The proposed laws add greater specificity to the hypotheses concerning underlying states, but they can be distinct from views concerning the agents that act in accordance with these laws. Thus, scientists can defend these proposed laws while remaining agnostic concerning the underlying agents. The primary sense of the term ‘theory’ Faraday employs in the “Historical Sketch” refers to a comprehensive explanation of the underlying agents responsible for producing the phenomena and the processes by which they bring about these effects. As such, Faraday focuses primarily on Ampère's views, which were “the most extensive and precise, and [had] been tested by the application of facts, and calculation very far beyond the rest. Indeed it is these alone among all those that have been given to the public, which deserve, if any do, the title of A Theory” (Reference Faraday1821–22, 111).

Faraday's theoretical diffidence in the “Historical Sketch” primarily concerns speculation about the underlying agencies responsible for producing the phenomena and the underlying states of the elements within an experimental system. For instance, in his discussion of the voltaic apparatus Oërsted employed in his initial experiments, Faraday observes that many assumed that the electricity produced by this device was a material fluid (or fluids) traversing the wire in the form of currents. But Faraday (Reference Faraday1821–22, 196) argues, “There are many arguments in favour of the materiality of electricity, and but few against it; but still it is only a supposition; and it will be as well to remember while pursuing the subject of electro-magnetism, that we have no proof of the materiality of electricity, or of the existence of any current through the wire.” And Faraday criticizes Ampère's assumptions concerning the underlying physical state of a conducting wire as it carries an electrical current. Given its centrality to Ampère's theoretical views, Faraday (Reference Faraday1821–22, 112) argues that one should “be allowed to expect that a very clear description will first be offered of it.” But, according to Faraday, Ampère fails to provide a clear and precise idea of this underlying state, which leaves Ampère's theory obscure at best. He writes,

Faraday's discussion of Ampère's views also illuminates a central problem for theoretical speculation: it can bias or obscure the understanding of experimental results. Ampère had proposed that some of his experimental results could be understood in terms of the activity of an electrodynamic electrical current, but Faraday notes that Ampère was “forced by his experiments, and the view he had taken of them, to conclude, that all the attractions, whether excited by two wires, a wire and a magnet, or two magnets, were purely electrical, and, in fine, that all magnetic phenomena are occasioned by electrical currents” (Reference Faraday1821–22, 114). Ampère's electrodynamic theory, while comprehensive in scope and substantially supported by experimental evidence, forced him to understand experimental effects in accordance with his theoretical views and potentially biased or obscured the findings disclosed by experiment. Ampère's theoretical views may have prevented a more apt description of experimental effects that could have generated a distinct and perhaps more adequate understanding of the phenomena and its causes. Thus, Faraday's decision to restrict his discussion in the first two sections of the “Historical Sketch” to the facts as they were discovered rather than to theories attached to these facts reflects his concern that higher-level explanatory claims may obscure or bias one's understanding of the phenomena produced by experimentation.

The goal of the next two sections is to show how Faraday constructs a historical account of this varied research and gives an account of the “facts as they were discovered” independent of the theories attached to them. I argue that Faraday accomplishes this task in two stages. First, his account suggests a distinction between theory-guided and exploratory experiments. The exploratory nature of many of the experiments in the early history of electromagnetism establishes that many experiments cannot be characterized as motivated or guided by problematic theoretical proposals of the kind specified above. Second, he illustrates the power of several dissociative methods of description that enable experimentalists to detach experimental effects from generating theoretical frameworks. Thus, even if theoretical speculation directs a particular experimental inquiry, this does not entail understanding experimental phenomena in terms of the framework theory.

3. Theory-Guided and Exploratory Experimentation

In Section 2 of the “Historical Sketch,” Faraday outlines at least 35 distinct sets of experiments in varying levels of detail. His account follows a chronological order and illuminates the diverse attempts to confirm, replicate, and extend the understanding of electromagnetic phenomena. Although the primary goal is to describe the discoveries, he also describes how these experiments enabled and motivated new avenues of research. For instance, some of these experiments led to the production of new scientific instruments that would prove to be extremely useful in subsequent research. Faraday places significant weight on Ampère's experimental research in this discussion and, as such, often provides minute descriptions of his experiments and the ways in which he sought to bring experimental results within the scope of his theoretical framework. Faraday's descriptions of other experimental processes and results are not nearly as detailed; in some instances, he merely states that a particular set of experiments confirmed the production of particular experimental effects.

Nonetheless, Faraday describes a variety of experimental processes, some of which are grounded in particular theoretical predictions and some of which are exploratory. He discusses experimental research of the following kinds:

1. 1. Experiments to confirm theoretical speculations concerning either (i) the underlying productive causes of phenomena or (ii) the underlying states of elements in the experimental system. These can include experiments to fit novel, unpredicted phenomena within the scope of the generating theory and experiments to create new instruments suggested by the generating theory.

2. 2. Experiments to establish a law describing the observed regularity of a specific set of phenomena.

3. 3. Experiments to contextualize new experimental effects in relation to other known phenomena.

4. 4. Experiments to confirm experimental results.

5. 5. Experiments on scientific instruments, including experiments to determine the typical functioning of the instrument and experiments to determine how alterations to an apparatus alter the typical effects produced by the apparatus.

6. 6. Experiments with new scientific instruments, including experiments to reproduce known experimental results, experiments to make experimental effects easier to manifest, and experiments to magnify experimental effects.

7. 7. Experiments to locate sources of experimental failure (e.g., the failure to produce some experimental effect that prior experimentation has established as a genuine effect).

This taxonomy of experimental processes shows that many forms of experiments are not performed with any explicit theoretical motivations in mind. In fact, the overarching aim of experiments of types 2–7 is to develop a clear understanding of particular experimental effects and their relations to other phenomena. Achieving this goal requires (i) establishing or stabilizing an experimental effect such that one can learn about its essential nature across wide variations in experimental parameters, (ii) contextualizing it in relation to other phenomena such that one can learn the various connections among phenomena, (iii) determining how to use these results to learn more about related phenomena, and (iv) learning from experimental failures so that one can counteract the potential confounding factors operative in experimental systems. For this reason, I call these kinds of experiments exploratory experiments. Given that experiments of type 1 involve the tests to confirm and expand the scope of a theoretical framework, I call these theory-guided experiments.

Exploratory experiments are not dependent on any of the problematic kinds of theoretical views that Faraday critiques in the “Historical Sketch.” This is not to say that exploratory experiments are independent of any form of theoretical or conceptual insight. Rather, experiments of this sort are directed at learning about phenomena as such. It is a separate matter whether one can employ this understanding to construct an adequate explanatory theory of the causes of phenomena or bring the results to bear on an existing theory within a domain. Often exploratory experimentation is predicated on provisional modeling of phenomena, but this mode of conceptualizing experimental effects is distinct from the forms of theoretical speculation that Faraday identified as problematic in the “Historical Sketch.”

Thus, the function of experimentation in generating evidence is not restricted to its capacity to generate effects that confirm or disconfirm a particular theoretical framework. Experimentation also generates evidence that deepens one's understanding of phenomena independent of any particular theoretical account of causal processes or physical states underlying the observed phenomena. The function of exploratory experimentation is separable from its capacity to substantiate more comprehensive explanatory theories. And precipitous attempts to situate experimental effects in the context of these theories can impede the process of deepening a lower-level experimental understanding of phenomena.

4. Theory-Guided Experimentation and the Dissociative Methods of Description

But it was much more problematic for Faraday to provide a historical account of the facts disclosed by theory-guided experimentation because these experimental processes cannot be understood apart from a specific theoretical framework. The commitments implicit in theory-guided experimentation can affect all aspects of the experimental process, from the observation of results, to the understanding of various elements of the experimental system, to the description and interpretation of experimental effects. Theoretical commitments can influence observation by directing one's attention toward particular experimental results, by structuring one's observations according to the conceptual resources of the generating theory, or by causing one to neglect potentially disconfirming experimental results in spite of their presence among the results of an experiment. Moreover, a theoretical understanding of the underlying states of various elements of an experimental system can determine one's interpretation of the kind of results typically produced by experimental instruments. This may cause one to perceive atypical results as artifacts of the experimental system rather than as genuine effects worth further scrutiny. And one's theoretical commitments can affect the description and understanding of experimental results. This is most obvious in cases where, to describe experimental effects, one employs terminology that has meaning only within the theoretical context that engendered the experimental process.

It is important to note, then, how Faraday reconstructs Oërsted's and Ampère's theory-guided experiments such that experimental results are dissociated from their generating theoretical frameworks. Analyzing Faraday's discussion of some of Ampère's theory-guided experimentation will provide an instructive illustration of these dissociative methods. Ampère's first important discovery was the fact that two parallel conducting wires carrying voltaic currents exhibit the phenomena of attraction and repulsion depending on the direction of the currents flowing through the conducting wires or, as Faraday prefers to put it, depending on whether one connects the extremities of the wires to the voltaic apparatus in the same or opposing directions. Faraday (Reference Faraday1821–22, 275) writes, “This discovery seemed to liberate the phenomena of magnetism from any peculiar power resident in the magnet, and to prove its production by electricity alone.” Ampère's substitution of an additional wire for Oërsted's magnetic compass needle resulted in the same set of experimental effects, and thus, “It was perfectly correct to consider these powers and actions as magnetical; so that it became evident that magnetism could be exerted independent of magnets, as they are usually called, and of any of the means of excitation employed, but wholly by electricity, and in any good electrically conducting medium” (Faraday Reference Faraday1821–22, 275).

According to Faraday, this experimental discovery establishes that magnetic effects can be produced by purely electrical causes. But these facts did not confirm Ampère's theoretical views because the attractions and repulsions were contrary to common magnetic attractions and repulsions; in this discovery, similarly directed electrical currents attracted each other and dissimilarly directed electrical currents repelled each other. Nonetheless, if one's theoretical framework requires that magnetic effects be fully explicable in terms of electrical causes, then this discovery constitutes a first step toward confirmation of these views by demonstrating the possibility of bringing about magnetic effects by electrical causes alone.

An important part of Faraday's description of these experiments, however, is his careful articulation of Ampère's experimental results. Notice that Faraday states (i) that these experiments “seemed” to liberate the effects from any peculiar power of magnets themselves and (ii) that Ampère was justified in considering the attractions and repulsions as magnetic even though they were not occasioned by magnets. And Faraday prefers to describe the attractions and repulsions without reference to the existence of currents flowing through the conducting wires. Faraday's reticence shows that he accepts Ampère's discoveries without committing himself to Ampère's theoretical views. Ampère's discoveries, at most, commit one to the claim that electricity can bring about magnetic effects, not that all magnetic effects are, in reality, brought about by electrical causes. This strategy of articulating experimental results cautiously without describing them in ways that would confirm the theoretical framework at issue is one method for dissociating experimental results from their enabling theory.

Faraday's discussion of François Arago's and Ampère's experiments on helically shaped conducting wires will help elucidate another of these dissociative methods of description. Faraday writes,

Although Faraday acknowledges that the experiments designed in consequence of Ampère's theory confirmed his theoretical views, Faraday's descriptions of these experiments tend to focus on the scientific instruments Ampère constructed in order to test these views. Thus, in regard to the experiments demonstrating that helically shaped wires created stronger magnetic effects on initially nonmagnetized bodies, Faraday describes at length the manner in which Arago and Ampère constructed various kinds of helically shaped wires and the resulting effects on needles placed within the axes of these wires.

And Faraday describes the magnetic effects resulting from these experiments in accordance with his own models, or construals, of Oërsted's initial discovery.Footnote ⁵ These models were phenomenological descriptions of the deflection of a magnetic needle when placed within close proximity to a conducting wire and, as such, made no reference to Ampère's electrodynamic currents. Thus, Faraday demonstrated that one can describe these findings in terms of provisional models that are independent of Ampère's theoretical views. The second dissociative method of description, then, involves conceptualizing experimental effects in terms of models independent of a generating theoretical framework.

Earlier I noted that Faraday's critique of Ampère's reliance on a theoretical understanding of electrical currents was, in part, grounded on his concerns with theoretically biased descriptions of experimental results. Ampère's theory was obscured by the fact that he failed to provide a clear account of the state of a conducting wire produced by the current as it traverses the length of the wire. But Faraday contends that Ampère's discoveries would have been better expressed in terms of experimentally grounded conceptions of the “powers” operative in the instruments employed in the experiment. If Ampère had articulated his results thus, he would not have had the additional burden of defending the theoretical interpretation of his results in terms of electrical currents. The description of Ampère's results according to Faraday's models demonstrates that one can avoid theoretically infused descriptions of results. Faraday argues that one should describe experimental results only in terms of models directly grounded on experimental results and not by appealing to underlying states for which there is no independent evidence. This is Faraday's third strategy for ensuring that descriptions of experimental effects are independent of the theoretical origins of the experiments leading to these discoveries.

In this section, I have identified three strategies Faraday elucidates to dissociate experimental results from their framework theory-dependence. First, one can employ terminology that avoids any commitment to understanding the results in terms of central explanatory posits within the enabling framework theory. Second, one can show how a particular discovery can be described in terms of models independent of the theoretical framework guiding the research. Third, one should model experimental results in accordance with phenomenological models rather than models of unobservable underlying states of the elements of one's experimental system. These dissociative methods allow one to counteract the potentially obfuscating and biasing effects of describing experimental results in terms of a generating theoretical framework.

5. Conclusion

Faraday's historical reconstruction of the experimental and theoretical developments in the science of electromagnetism provides greater insight into the epistemic significance of experiment. His arrangement of the material discoveries and products of experimental research into a systematic form provides an illuminating account of the manner in which exploratory experimentation enables one to understand experimental effects apart from a theoretical explanation of its underlying productive causes. Nonetheless, Faraday understands that theory can play a central role in generating experimental research that leads to the discovery of significant facts and the creation of scientific instruments. The primary importance of theory-guided experiments is not the confirmation that successful testing provides. Rather, these experiments (i) disclose phenomena that any theoretical proposal must explain, (ii) bring to light novel phenomena requiring subsequent study and analysis, and (iii) promote the construction and use of scientific instruments that become an essential aid in experimental investigation. Faraday's focus is on the products of theory-guided experimentation that one can detach from the generating theoretical framework.

Breaking the framework theory-dependence of theory-guided experimentation involves methods relating to the observation, description, and modeling of experimental results. With respect to description, one must resist the immediate inclination to describe effects in terminology that derives its meaning from the framework theory. One should instead describe experimental results in a manner closely tied to experimental observation. Likewise, provisional models of experimental effects should be formulated in close association with experimental observation, with avoidance of any theoretical constructs that appeal to unobserved states of substances that could act as causes of phenomena.

Exploratory experiments are not necessarily free from all forms of theory-dependence because conceptual insight is often an essential element in the formation of a suitable model for experimental effects.Footnote ⁶ Thus, Faraday's approach to the systematic detachment of experimental results from framework theory-dependence does not necessarily provide grounds to dissociate experimental results from all forms of theory-dependence. But one can adopt methods employed in exploratory experiments, such as the systematic variation of experimental parameters, to mitigate the potentially biasing effects of misleading models. The systematic variation of experimental parameters and the attempts to bring about experimental results within experimental systems dependent on divergent assumptions ensure the stability of experimental results across possible changes in theoretical framework.Footnote ⁷ Often it is the stability of these results that solidifies specific models into the proper interpretations of experimental results.

So there is an interplay between theory and experiment at this lower level of investigation independent of theoretical explanation of productive causes and the material nature of objects under investigation. And one can avoid obvious bias in the modeling of experimental results by revising these models in light of experimental tests. Thus, while experimental results are not completely independent of conceptual insight, the methods involved in exploratory experimentation lessen their dependence on particular conceptual models. These considerations lend credence to Faraday's conception of theories built on stable experimental results while acknowledging that experimentation itself cannot produce evidence independent of conceptual insight.