Abstract

In De magnete (1600), Gilbert conceptualizes magnetism as primarily disponent, that is, as primarily aligning or ordering magnetic bodies with respect to each other. The conceptualization of magnetism as disponent replaces that of magnetism as attraction. The focus on magnetism as disponent is a consequence of Gilbert’s treatment of magnetic motions and magnetic power. A reading of Gilbert’s conceptualization of magnetism as disponent makes sense of certain argumentative and investigative choices in De magnete. Among them is what has otherwise been seen as a baseless claim in support of Gilbert’s alleged Copernicanism: that certain magnetic phenomena are rotational, and that the Earth thus rotates because it is a giant magnet.

Introduction

In A Treatise of Artificial Magnets, John Michell observes,

Not being aware of this property [i.e. the equality of attraction and repulsion], he [Gilbert] concluded from some experiments he had made, not very irationally [sic], that the Needle was not attracted by the magnet, but turned into its position by, what he calls, a disponent virtue […]. (Michell 1750, p. 17; my emphasis)

For Michell, the disponent virtue (Lat.: disponens vigor)1 is the underlying cause of magnetic phenomena in Gilbert’s treatment. He is not alone. Ridley (1613) and Carpenter (1635) also read Gilbert’s magnetism as (at least partially) positing a disponent virtue.2 I agree. In this paper, I show what it would entail to treat Gilbertian magnetism as disponent.

To the extent that it is disponent, magnetism arranges and aligns magnetic bodies into place (where the place of each body is specified relative to a different magnetic body). This is not inconsequential: magnetism as disponent is a departure from the established conception of magnetism as attraction. Taking magnetism as disponent informs both the way in which Gilbert accounts for magnetic phenomena (as specific movements into alignments) and the way Gilbert went about experimentally investigating these phenomena (with a particular focus on mapping the relative positions and directions of magnetic bodies).

This paper is not concerned with the meanings Gilbert attributes to the term “disponens vigor” (or its variations) itself, nor do I provide an intellectual-historical treatment of it. Instead, I take on Gilbert’s term as an umbrella term for a conceptualization of magnetism that (1) moves away from the formerly dominant treatment of magnetism as attraction, emphasizing instead (2) the relationist/interactionist and organizational nature of magnetism. Gilbert captures this conceptualization by showing that each magnetic body’s action manifests strictly relative to another magnetic body. Articulating what this entails and how it works epistemically and heuristically is the goal of this paper. There are thus two interrelated sides to my argument: (1) magnetism understood as disponent is central and (2) magnetism understood as disponent guides Gilbert’s theoretical and investigative practices.

Despite the diverse roles it plays in Gilbertian magnetic philosophy, a relational–organizational treatment of magnetism has been largely neglected in the scholarship. This paper supplements this by offering an interpretation of Gilbert’s treatment of magnetism as disponent and of how it shaped his local arguments, focusing especially on how magnetism was taken to substantiate the Copernican universe in general and the Earth’s axial rotation in particular.3 Gilbert’s argument goes as follows: the Earth is a magnet, and magnets rotate; therefore, the Earth rotates because it is a giant magnet. Historians have puzzled over the claim that magnets rotate, and have largely concluded that Gilbert has little evidence to offer in support of the claim (Henry 2001; Miller 2014). I show that Gilbert's assent to magnetic rotation is (at least, partially) a consequence of both the treatment of magnetism as disponent and the manner of investigating this. If my interpretation holds, we need to consider the disponent conceptualization of magnetism in order to understand further appropriations of the Gilbertian framework, both in the short-lived magnetic philosophy itself4 and in the work of others influenced by Gilbert, such as Stevin’s The Heavenly Motions (1608), Kepler’s The New Astronomy (1609), or even Descartes’s explanation of magnetism in the Principles of Philosophy (1644).

My analysis also contributes to debates centered around Gilbert’s experimentalism. The literature has focused on either Gilbert’s sources for his experimentalism or on critical evaluations of the extent to which he was a pioneer of the scientific method. I leave the latter to one side (because I am skeptical about both the possibility of there being any unitary scientific method and the value of ascribing one). On the question of Gilbert’s sources, Zilsel (1941) (following his Marxist account of science) argues that Gilbert’s experimentation follows the mechanical arts, while Henry (2001) concludes that it is more natural to place Gilbert in the natural magic tradition, and Gaukroger (2006) suggests that De magnete is in line with the natural-historical tradition. In establishing intellectual lineages, it is of little use to have an account of the many roles that Gilbert’s retrospectively reported experiments and experimental results played in the articulation of his arguments. The current paper shows the dynamic at work between Gilbert’s conceptualization of magnetism and his experimental practices.

I begin by showing that conceptual reform in magnetism was an explicit goal of Gilbert’s treatise (Section 2). This was necessary because accounting for magnetism through attraction was unsuitable (Section 3). Magnetism as primarily attractive is replaced with magnetism as disponent, which takes magnetism to primarily act with respect to other magnetic bodies, and align or organize magnetic bodies into privileged relative positions. Each magnetic phenomenon can be accountable for a different type of alignment of magnetic bodies (Section 4). Magnetism as disponent is suggested by the investigative practice; further, taking magnetism to be disponent affects both the classification of what one takes a magnetic phenomenon to be and how one goes about treating it in the investigation, as the case of magnetic rotation shows (Section 5). Before entering the main body of the paper, though, I provide a broad overview of Gilbert’s “magnetic philosophy” since De magnete is not a widely read treatise today.

1. Gilbert’s Magnetic Philosophy

The celebrated central claim of De magnete is that the Earth is a giant magnet. It is not that the Earth contains some magnetic substance: the Earth itself is magnetic. The status of this claim, whether it is a demonstration derived empirically, or whether it is simply posited, is hard to establish. There are several reasons for this. 1) We do not have the notes from Gilbert’s experiments, and we know rather little about the actual composition of De magnete. 2) The order of presentation of the arguments in De magnete is most definitely a retrospective account, but one intended to provoke assent to its claims: the evidential bias is to be expected. 3) Conceivably, there is a third option (and there might be others): by taking the Earth’s magnetic nature as a likely scenario, Gilbert reworked his findings. This third option would be a synthesis of the first two.

For the purposes of this paper, what is important to note is that De magnete’s arguments do indeed build towards showing the cogency of the claims that the Earth is a giant magnet and that its properties as magnetic can be studied locally with loadstones and iron, as the two are closest in nature to the Earth (Gilbert 1600, p. 41). For the terrestrial region, Gilbert adopts a single-element theory of matter: the elemental pure earth.5 But the pure earth is to be found untainted only in the depths of the Earth, while the bodies at the Earth’s surface layers consist of impure earth, corrupted by celestial and atmospheric influences (Gilbert 1600, p. 41). The elemental earth is magnetic, and amongst the bodies known to us, the loadstone and iron are closest in nature to this pure element, since, “[…] they [i.e., loadstone and iron] both are true and intimate parts of the earth and as such retain the prime natural properties of mutually attracting, of moving and of disposing themselves toward the position of the world, and of the terrestrial globe; which properties they also impart to each other, and increase, confirm, receive, and retain each other’s forces” (Gilbert 1600, pp. 37–8). This is why loadstone and iron can be used as proxies for studying the nature and motions of the Earth itself.

Gilbert’s matter theory and his commitment to the Aristotelian principle that “the natural movements of the whole and of the parts are alike” then grounds what I term an “extrapolative experimental methodology”, aimed at thoroughly investigating in contrived conditions various magnetic interactions, in which results established by investigating loadstones and iron are extrapolated to the Earth itself (Gilbert 1600, p. 224). And Gilbert rallies plentiful experiments to this end. The myriad of experiments are performed with a few devices, in particular (a) the terrella (“miniature Earth”—a device obtained by shaping a loadstone into a sphere, with the magnetic poles, magnetic equator and (at least some) magnetic meridians marked on its surface) and (b) the versorium (a device that allows a (ferrous) metal needle suspended on a fulcrum to move freely on the horizontal axis and/or the vertical axis).6 As an important aside, the extrapolation of results from the terrella to the Earth explains why Gilbert posited an axial, rather than a tilted, magnetic dipole.7

Following this methodology, Gilbert classifies the magnetic phenomena (or magnetic movements) as follows: (1) magnetic coition (the coming together of two magnetic bodies according to the rule that opposite poles unite); (2) magnetic direction “towards the poles of the Earth” (the movement into alignment of a compass needle with the magnetic poles); (3) magnetic variation (a deflection from the magnetic meridian); (4) declination (the deviation of a magnetized needle on the vertical plane); and (5) rotation (or circular motion) around an axis (Gilbert 1600, p. 46). These are the various ways in which magnetic bodies act upon and influence each other under determinate conditions. The further inference is that we encounter these magnetic phenomena precisely because magnetism acts as a disponent power (even coition occurs only after disposition; see section 5.4). De magnete is structured around their investigation.

2. The Conceptual Reform of the Magnetic Philosophy

Part of the argument of this paper is that treating magnetism as primarily disponent replaced treating magnetism as primarily attractional. Gilbert appears to be well aware of a need for such conceptual change in the investigation of magnetism. This section argues as much, by showing that Gilbert’s methodological remarks indicate that conceptual choices matter in the articulation of previously unknown things (Section 2.1) and that not all conceptual choices are equal; some are more appropriate than others (Section 2.2.). The claim is not merely that some terms are replaced with others, but that Gilbert was explicit about the need to introduce new concepts into magnetism in order to articulate the magnetic philosophy.

2.1. The Quest for “New and Unusual” Words

In the introduction to the treatise, Gilbert states,

Sometimes therefore we use new and unusual words, […] [so] that hidden things which have no name, never having been hitherto perceived, may be plainly and correctly enunciated. (Gilbert 1600, p. iii; my emphasis)

“New and unusual” words are introduced both to name something and to signpost its empirical disclosure.8 There are two sides to disclosing otherwise hidden things: procedural and descriptive. On the procedural side, by which the thing is made accessible in the investigative context, the disclosure of a new thing is fundamentally a process of differentiation, by which something is picked out as (at least) sufficiently regular and sufficiently significant to stand in a category of its own. The exact criteria for what constitutes something as novel will indubitably differ, but in this case, Gilbert makes at least one such criterion clear: the thing needs to have been perceived (Lat.: perspecta) (Gilbert 1600, p. ij). Perceptual differentiation is necessary for disclosing something. The perceptual differentiation needs to be articulated, and to this articulation corresponds the descriptive side, an activity by which those aspects that matter and those that do not are accounted for.9

As a simple example, when magnets come together to unite, Gilbert considers that their sizes and shapes are relevant and need to be accounted for, while their respective colors are not and do not. It might be warranted to bypass the variable of color, but the process of establishing through investigation which variables matter and how to account for them is essential for establishing what a thing is. A case in point in De magnete is the separation of bodies into two natural kinds: magnetic bodies and the “electricks.” The electricks are that class of bodies that attract other bodies materially (see Gilbert 1600, pp. 47–61), while magnetics influence each other without any material exchange. The term “electrum” is Latin for amber (amber being especially susceptible to the buildup of static charge), but Gilbert extends it to all bodies that show the same kind of material attraction that amber does, and in this way he engineers a new classification of bodies relative to whether or not they act by material exchange.10

The procedural and descriptive sides go hand in hand, and they cannot be fully separated (at least not in the initial stages). What the “hidden thing” is, and how it is best described, is entrenched in the investigative situation; without much practical expertise with the investigative situation, what is being communicated about the hidden thing might turn out to be mysterious or at least difficult to grasp. It usually takes extended work to turn a “hidden thing” into an independent, self-standing thing that might eventually be dragged to a different domain in a non- ad-hoc manner. Thus, although it might initially seem unproblematic, articulating a novel discovery through words is neither easy nor inconsequential. The descriptive choice matters.

2.2. Some Concepts Are Better “Fitted” than Others

The second passage where I take Gilbert to indicate that reconceptualization is necessary reads,

[…] afterward numerous subtleties, hitherto abstruse and unknown, hidden in obscurity, are to be laid open, and the causes of all these (by the unlocking of nature’s secrets) made evident, in their place, by fitting terms [verbis idoneis] and devices. (Gilbert 1600, p. 15; my emphasis)

The focus here is no longer on the novelty of terms, but rather on their appropriateness, their suitability. “Nature’s secrets” might be disclosed with the help of devices, but we also need “fitting words” to describe what was disclosed. We are not told what the criteria for “fittingness” are, but, given the commitment to experimentation, it is not far fetched to take it that whether it fits (what are taken to be) the relevant experimental findings is one criterion.11 Once a description is provided in such fitting terms, it directs the explanations one gives. For instance, Miller (2014, chap. 3) nicely shows the constitutive role of Gilbert’s description of magnetic bodies in terms of poles, meridians and equator in his explanations. I would add here that the conceptual choices also play a direct role in the investigative process: they afford certain possibilities of continuing the inquiry and set certain limits on how one is to think further about the phenomenon in question. Some possibilities of problematizing the phenomenon are shut down, while others are opened up, to the extent that the very selection of terms legislates how to go about the investigation. Shifting from attraction to coition is again a case in point (explicit examples of this are discussed in section 5).

Reconceptualization of magnetism is thus an explicit goal of De magnete. Gilbert is on the lookout for both “new and unheard-of” and “fitting” terms. The new and fitting terms influence the descriptions and explanations, but they also inform the investigation and drive it down certain paths, while closing off others. Of course, these are just some of the layers within the dynamic of the interaction between investigative practices, conceptualizations and research goals. Cases of how the disponent power shapes the trio of laboratory practices, descriptions, and explanations will be discussed in section 5, but not before showing what Gilbert’s treatment of magnetism replaces (section 3) and addressing in greater detail what the disponent power is, and how it is supposed to work (section 4).

3. The Problems of Magnetism as Attraction

In the previous section, I argued that Gilbert took magnetic philosophy to be in need of new and suitable concepts. Book 1 of De magnete claims that previous studies of magnetism were flawed. Of particular interest is Gilbert’s discontent with attraction-based accounts of magnetism. Attraction was problematic because it localized magnetism (Section 2.1). And localization was problematic because Gilbert rejected accounts of motion based on a substantive notion of place (Section 2.2).

3.1. Localizing Magnetism

To Gilbert, attraction involves force: “[…] there seems to be force applied where there is attraction and an imperious violence dominates” (Gilbert 1600, p. 61). If applied to magnetism, the familiar class of phenomena used as exemplars are dragging cases: in the same way a horse drags a cart, magnets drag to themselves pieces of iron. Such treatment of magnetism was familiar at the time, and is in line with the Aristotelian agent–patient conception of change.12 Broadly, while both the agent and the patient are oriented towards the same definite end (the realization of a particular state in the patient), the agent is the only source of the change. In this case, the loadstone is the causal agent and the iron the patient. Force then in attraction entails causal asymmetry.

Because it dictates which criteria of the investigative practice are appropriate and which are not, treating magnetism as attraction, asymmetrically, affects both how one investigates magnetic phenomena and how one accounts for them. For instance, under this asymmetric understanding of magnetism, the movement of a compass needle will be explained in terms of the attraction of the passive needle towards an active source of magnetic power. With this frame of reference guiding the explanations and investigations, the standard investigative practice was to search for the actual physical location of the magnetic power at the terrestrial scale.

Through cartographical methods, triangulation of observations, or mere speculation, many tried to locate the purported source of the Earth’s magnetism: Peregrinus claimed that the magnetic poles correspond to the celestial poles, Cardano deemed the position of the magnetic pole to be located in the Polestar, while Martin Cortes concluded that the origin of the magnetic force lies somewhere outside the outermost sphere. In the sixteenth century, it became somewhat fashionable to place the magnetic source somewhere on the surface of Earth, in a magnetic mountain in the Red Sea, or in a magnetic island, and so on.13 Usually, because such investigations of magnetism made measurements almost exclusively in the northern hemisphere, the geomagnetic south pole rarely played a role in these accounts. Magnetism was usually claimed to be located in the north.

In The New Attractive (1581), Robert Norman provided an experimental refutation of an asymmetrical geomagnetic source.14 Norman had shown that a magnetized iron needle attached to a piece of cork and floated below the surface of the water in a vessel remains in equilibrium. It is not pulled in any direction. The conclusion of his series of experimental observations was that there is no source of attracting power. Had there been a source of attraction in the heavens, the magnetized needle would have moved upwards; had there been one in the Earth, it would have moved downwards. Gilbert was satisfied with Norman’s experimental proofs (Gilbert 1600, pp. 67, 125), but was less inclined towards Norman’s alternative explanation. Norman replaced the attractive framework with one that posited that the magnetized needle “respects” a “respective poynt” (Norman 1581, p. 26). Gilbert was concerned that Norman’s account relies on the positing of an occult quality, on account of the non-manifest nature of his respective point (Gilbert 1600, p. 162). There is further reason for Gilbert to resist Norman’s respective point: although it rejects attraction, it is still an asymmetrical conception. Norman’s account did not provide a viable alternative insofar as it respected the same underlying logic as the attractive alternative—it localized magnetism to a particular source. It is now the respective point that becomes the source of the magnetic action. In other words, magnetism is still asymmetrical! It has a causal and physical source in some particular, privileged location.

3.2. The Emptiness of Place

We have seen that Gilbert rejects the asymmetrical conception of magnetism because it ultimately involves localization (both causal and physical). Gilbert has experimental reasons to reject this asymmetrical account; he also rejects (Aristotelian) accounts of motion that require the positing of a “natural place,” as a self-standing metaphysical entity, in order to account for motion of bodies. This rejection makes a conception of magnetism that relies on localization impracticable.

In De mundo (published posthumously in 1651), Gilbert vehemently rejects a substantive account of place15:

Place neither operates in nor rules over the nature of things in order to make bodies rest or move. […] Place is neither a being nor an efficient cause: the power frequently flows into the contents from the surrounding bodies.16 (Gilbert 1651, p. 61)

The point here is the rejection of “true place” as a metaphysical entity (neither a being nor an efficient cause) against which a body can be judged to be in motion or at rest. Gilbert rejects absolute place and affirms relativity of motion. His position accounts for motion relative to (a) the power that flows from (b) the surrounding bodies. I read this as suggesting that the attribution of motion (and rest) to a body is possible only relative to the other bodies with which it interacts. The interaction of multiple magnetic bodies thus must be a precondition for any magnetic motion: it is in relation to them that the motion occurs; it is from them that the “power” flows. And in a magnetic interaction, all the bodies involved are causally active. Hence magnetism must be symmetrical for Gilbert.17

3.3. From Attraction to Coition

That what is at stake is not simply a terminological shift away from attraction, but a conceptual shift from an asymmetrical source-based conception of magnetism to a perfectly symmetrical and relational conception, is further clarified by the case of magnetic coition. To move away from attraction, Gilbert describes the observation that magnetic bodies unite as magnetic coition (Lat.: coitio). Coition is a concept introduced to capture the causal symmetry, or (to put it in terms closer to Gilbert’s vocabulary) the mutuality of magnetic action. Coition is, “a primary running together,” a mutual action of “the loadstone and of the iron, not an action of one […] an action of each towards unity by the conjoint action and συνεντελέχεια of both” (Gilbert 1600, p. 62).18 In magnetic interaction, then, iron does not receive the power from the loadstone; nor does it respond magnetically after the loadstone’s action, nor as a consequence of the loadstone’s action. The claim is thus not that it is empirically undecidable which object is the agent and which the patient, but that both loadstone and iron causally contribute to magnetic change (or motion). That this is so, Gilbert attempts to show experimentally, but his experiments in themselves do not definitively prove coition: the experimental interpretation of the results corroborates the prior commitment that both magnetic bodies are agents.

Accounts of magnetism as attraction are given up in favor of a symmetrical/mutual conceptualization of magnetic change that distributes the causal contribution between all the magnetic bodies involved in the interaction. The interaction thus has explanatory priority over its parts. And, in magnetism, it is such interactions that account for magnetic phenomena. These interactions are further accounted for by treating magnetism as disponent.

4. Conceptualizing Magnetism as Disponent

The previous section showed Gilbert’s commitment to the mutuality of magnetism. This section shows what it entails to conceive of magnetism as disponent. I first show that Gilbert prioritizes magnetic interaction as a system over the individual magnetic bodies that contribute to the interaction. I use Gilbert’s disponent power to refer to the manifestation of magnetic power that aligns magnetic bodies relative to each other in magnetic interactions (section 4.1). I briefly show (section 4.2) that Gilbert accounts for the magnetic phenomena we observe precisely because magnetism is (primarily) disponent.19

4.1. From Magnetic Power to the Disponent Power

The magnetic power is the innermost nature of magnetic bodies. It is not a part, nor is the magnetic power superadded to the earth-matter. Put simply, the magnetic power is the necessary and sufficient condition for a body to be a magnet.20 As a power, it is active. If active, it seems reasonable to attribute real agency (and thus change and/or motion) to each magnetic body, which, in return, makes each and every magnetic body the productive cause of a given observable magnetic effect. On my reading, however, Gilbert disagrees. In the movement of coition, he is explicit about the fact that both bodies are mutually productive (Gilbert 1600, p. 62). Both bodies co-participate in magnetic actions. But this seems paradoxical, since a magnetic body by itself is supposed to have magnetic power. How do we reconcile this tension? My suggestion is that in order to solve the tension we need to show that (a) in the empirical investigation, the system of magnetic bodies has priority over each individual magnetic body, and (b) there is a conceptual distinction between the magnetic power and its manifestation (as either disponent or as coition, or others).21

Let’s start with the latter claim. It is undeniable that the magnetic force is inherent in each and every bit of magnetic matter (Gilbert 1600, pp. 65–71). As such, a magnetic body’s agency ultimately rests on its magnetic power. However, other than positing its existence, Gilbert can do little with magnetic power in itself. A non-metaphysical, but conceptual, distinction is in place between the magnetic power as a power, or that which gives the power to act, and the power’s manifestation or expression in given circumstances, or that in which the action is regulated and limited. As I read it, the magnetic power can manifest itself in various ways; there is a qualitative difference between its manifestation in magnetic coition and its manifestation in declination (see for instance section 5.4 below), or its manifestation as a celestial body.22 If so, then there is a distinction between the magnetic force itself and its expression as a disponent power.23 Aided by empirical investigations, Gilbert concludes that magnetic power qua physical—i.e., as analyzed through the magnetic phenomena—is primarily disponent, in that it directs and aligns the magnetic bodies by turning them into relative positions.24

The manifestation of the magnetic power as disponent is primarily the manifestation we encounter in magnetic phenomena. What seems to be at stake when Gilbert claims that magnetic bodies dispose, given how he makes use of the notion of “disponent” (and its correlates), is that magnetic bodies fall into order. For magnetic phenomena the order is spatial and is specifiable relative to the surrounding bodies. In some cases, the spatial order is one of relative positioning, while, under some circumstances (i.e., shorter distances and higher magnetic strength), it is translational movement such that the magnetic bodies unite (see section 5.4). It is the process of arranging, and the arrangement itself, relative to the surrounding bodies that takes priority and makes magnetism fundamentally interactionist.25 For a single body, magnetic power makes no physical difference precisely because a single magnetic body in isolation cannot constitute a magnetic phenomenon for Gilbert: it is a given that magnetic phenomena are types of interactions of magnetic bodies. It thus makes no sense to talk about magnetic phenomena of a single body.26 Magnetic phenomena occur in a system of magnetic bodies, not in isolation. Magnetism as a physical property is analyzable only in the context of a system of (two or more) magnetic bodies.

Now, Gilbert likens magnetic power to a soul: magnetic power is “animate, or imitates life” (see esp., Gilbert 1600, pp. 208–10). This might make it seem as though animism about magnetism explains magnetic activity. However, whether or not we take Gilbert’s animism seriously, its mere positing would hardly constitute an argument.27 What matters is whether it makes any difference explanatorily and phenomenologically. The animism appears to be of very little use to that end, given that Gilbert is interested in how magnetic power operates. Consequently, the position I advocate for in this paper is indifferent to how one might want to interpret the animist doctrine of magnetism.

For Gilbert, magnetism is understood through its effects: it is an empirical domain of investigation. And the core of Gilbert’s empirical theory of magnetism is that magnetic bodies do not attract—they align; they orient in space relative to each other just because they are magnetic. They orient because the “magnetical motion is one of arrangement [Lat: dispositionis] and conformation [Lat: conformationis]” (Gilbert 1600, p. 60), and by this motion, “[…] magneticks are disposed [Lat.: disponuntur] and turned [Lat.: convertuntur] and combine with magneticks in proportion as the parts facing and adjoined unite their forces together” (Gilbert 1600, p. 97).

Gilbert makes a distinction between two magnetic “orbs” (volumes within which magnetic power acts): the orb of virtue and of the orb of coition. For Gilbert, magnetism acts at a distance, but not every relative distance amounts to an interaction of magnetic bodies. Two magnetic bodies interact only when they are within an orb of virtue (Lat. orbis virtutis).28 Somewhat paradoxically, however, Gilbert appears to introduce the orb of virtue as a property of a single individual magnetic body, since it is “all that space through which the Virtue [i.e., the magnetic power] of any loadstone extends” (Gilbert 1600, vi).

This sounds as though the magnetic power’s manifestation is being considered with respect to an individual magnetic body. Notice, however, that beyond its definition, the orb of virtue only has the properties it does (its strength, shape, etc.) in the presence of a second body. That a second body is necessary for the empirical analysis of the orb of virtue becomes even clearer if we consider the distinction between the orb of virtue and the orb of coition. The orb of coition is “all that space through which the smallest magnetick is moved by the loadstone” (Gilbert 1600, vi). Unlike the magnetic motions accounted for exclusively by the disponent power, coition presupposes displacement of the magnetic bodies not only relative to their respective positions, but also in their relative distances, since coition is effectuated by the actual coming together of the magnetic bodies involved. The displacement in relative distances takes place within the orb of coition whereas the alignment of bodies happens in the orb of virtue (which contains the orb of coition). Moreover, relative alignment is a precondition for coition, since the effectuation of coition is possible only after north–south alignment; that is, in order for bodies to unite, they first need to align (dispose).

Before going into what it entails to understand magnetism as disponent, I would like to briefly deal here with an obvious objection: the claim that the disponent power is an occult quality. My reply is not independent of what one takes an occult quality to be. If an occult quality is something insensible, which is what Gilbert seems to have taken them to be (see Gilbert 1600, pp. 162, 207), then the disponent power is no occult quality. Here is why: (1) Gilbert’s frequent use of analogies between his inquiry and the passage from darkness to light suggests that one of the overarching goals of the treatise is to show that magnetism should not be classified as an occult quality because it is made manifest (accessible to the senses). The disponent power’s manifestation is made accessible to the senses (through experimental observations, but also through diagrams) in the mapping activities (see section 5). (2) And, to the extent that the criteria for successful accounts are settled by the actors themselves, Gilbert takes himself to have succeeded in making the magnetic power manifest and not transforming it into another occult quality. If Gilbert takes himself to have rendered the magnetic power itself manifest, then its manifestation through the disponent power certainly cannot be an occult quality in this sense.

If, on the other hand, one holds the view that an occult quality is the position that a vis or vigor has no explanatory power, one can still argue that the disponent power is not such a quality. First, Gilbert seems to take the disponent power to be something akin to the efficient agent of change and as such to have explanatory power (Gilbert 1600, p. 65). Second, the disponent power is explanatorily productive: it is used in, for instance, the construction of a nomograph for calculating latitude at sea. And third, the disponent power is not posited as one occult quality to explain this or that particular effect, but rather as at least one of the basic manifestations of the magnetic power present in all magnetic phenomena, thus unifying the magnetic phenomena under the same account.

4.2. Magnetic Movements through the Disponent Power

Understanding magnetism as disponent guides the way in which Gilbert identifies the magnetic phenomena, insofar as each is an effect of different relative alignments of magnetic bodies. As an example, let’s compare magnetic direction to magnetic declination. A case of magnetic direction is the movement of the magnetic compass needle so as to align its poles with the magnetic poles of the Earth: the north–south axis of the needle lines itself up with the north–south axis of the Earth. Any magnetic body has an inherent tendency to align its poles with the poles of another magnetic body, along a magnetic meridian. The movement of the magnetic body in magnetic direction is on the horizontal plane, along an axis perpendicular to both the horizon and the north–south axis of the moving body, and it is governed by the north–south axes of both magnetic bodies.29

Magnetic declination, on the other hand, involves a different alignment. In magnetic declination, a magnetic body turns along an axis perpendicular to its north–south axis and horizontal to the plane of the horizon to align with the other magnetic body. The particular alignment changes depending on the latitudinal position of the one body relative to the other. Just as we see magnetic direction in a compass’s north–south alignment, we see magnetic declination when the compass needle dips from the horizontal (a phenomenon first discovered by Norman). In this way, the movements of direction and declination are distinguished by their different types of alignment.

If an observable magnetic movement is not ultimately fully explained by the relative alignment of magnetic bodies, but by some other external cause, then it is classified as a “perverted movement”, that is a motion caused by factors external to the system of the magnetic bodies. Magnetic variation exemplifies this. For Gilbert, magnetic variation (at any particular location) is the degree of angular movement that a magnetized needle makes to the east or to the west of the magnetic meridian. It is a deviation from magnetic direction. Gilbert’s account of magnetism assumes an axial dipole, with no tilt: he takes Earth’s axis of rotation to be the same as its magnetic axis.30 Now, his study of the magnetic power around a terrella indicated a regular pattern of action. Measurements taken with respect to the poles of the Earth itself, however, indicated that the compass needle always deviates from the magnetic meridian. For Gilbert, the only plausible explanation to be found for this is the distribution of continental landmasses (plus the local effects that deposits of iron and/or loadstone would have). Hence, variation is a perverted movement.

Given the centrality of the relative alignments to his account, mapping exact alignment around magnetic bodies becomes an overarching goal of Gilbert’s experimental investigation. That mapping is a reliable practice is guaranteed by the regular, non-arbitrary nature of magnetic alignment.31 The mapping practice also leads Gilbert to indicate experimentally the magnetic rotation, as the next section will show.

5. From Magnetism as Alignment to Magnetic Rotation

I show that magnetism as disponent is the result of the investigative practice and that Gilbert’s conceptualization of the magnetic movements depends on taking magnetism to be disponent. The focus of the section is the movement of magnetic rotation—which is especially significant for Gilbert’s cosmology.

5.1. Magnetic Rotation

The “circular motion, or revolution” is one of the five magnetic movements we observe (Gilbert 1600, p. 46). It is a motion around an axis, for which magnetic bodies are “fitted” because they are bodies with poles, equator and meridians, natural bounds that are proper to circular motion. If the Earth is a magnet, and if magnets rotate, then magnetism explains the Earth’s diurnal rotation. At times, Gilbert’s claims about magnetic rotation can seem tenuous; the literature tends to take it that Gilbert does not have strong evidence for these claims (for instance Miller 2014, p. 79). Seen through the lens of the disponent power, however, I claim that Gilbert did have experimental reasons to claim magnetic rotation. Many experiments suggest partial rotation, and the investigation of magnetic declination shows full continuous rotation (see Section 5.2). It is in this latter context that Gilbert considers himself to have shown that, when one magnetic body orbits another, continuous axial rotation follows.

5.2. Magnetic Declination as Rotation

The universe of Mark Ridley, one of Gilbert’s close collaborators, agrees in major points with Gilbert’s cosmological model. Gilbert presents an ordered universe endowed with orientations. The orientations are of the north–south axes of the globes, and the order is that of the relative positions between the celestial bodies.32 It is the universe’s order and orientation that explains the positions of the globes, but the universe’s order is sustained by the astral nature of the globes (the Earth’s being magnetic), with the sun taking a privileged position. The sun seems to be the only source of causal asymmetry in Gilbert’s universe (Gilbert 1600, p. 224). Gilbert does not take a definite position on what sorts of alignments and distances are involved in the planetary system so that the planets are carried in their movements by their interaction with the Sun, but he does suggest that, whatever they might be, (at least) in Earth’s case magnetism can explain them, since magnetic bodies in declination show circular motion.

This section is concerned with showing that Gilbert had cogent arguments for plausibly attributing circular motion to magnets, and thus for maintaining that the Earth spins on its axis. These were arguments drawn from his investigation of declination or dip. Ridley appears to read Gilbert in the same way:

[…] the Earth having parts very fit for motion, as her Globious bodies with equator and parallels ready and apt to obey all internal virtue and power, […] which we have called elsewhere Magneticall, […] which kind of vertue is also demonstrated in the Earth, passing by the Meridians to the poles, as also experimentally the Inclinatory ring with his needle carried about the Earth upon a meridian circle, maketh this circular motion from the equator to the poles, and so back again […]. (Ridley 1613, p. 43)

“Declination” is Gilbert’s term for what we know today as magnetic dip. If one balances a needle on a pivot in perfect equilibrium but free to move on its axes, the needle, once magnetized, will position itself on both the horizontal and the vertical planes relative to the local lines of flux of the magnetic field. The phenomenon of vertical alignment is what Gilbert calls “declination”:

Declination is seen to be a motion of a needle, […] that motion arises in truth not from any motion from the horizon toward the centre of the earth, but from the turning of the whole magnetick body toward the whole of the earth […]. (Gilbert 1600, p. 185; my emphasis)

Every angular measurement of declination is a part of a rotation. This interpretation is supported by the investigation aimed at mapping out the declination at the surface and within the orb of virtue of a terrella.33

Gilbert reports many experiments in the chapters on magnetic declination, but one is especially relevant. The experiment proceeds as follows. Gilbert places a versorium at various points of latitude around the circumference of a terrella, starting at 0° (which he marks on the terrella’s equator) and moves it in a complete circuit.34 As the versorium changes latitude, it rotates. At any given point of latitude, the versorium “makes always one kind of angle”,35 and thus a declination–latitude correlation can be established. At the equator, the versorium will be horizontal (with respect to a tangent from the surface of the terrella), but by the time it has been moved to 45° latitude (from the equator), it will have dipped by around 100°. The movement will continue, reaching 180° at 90° latitude, and then 360° at 180° latitude.

In this way, the versorium will make two complete rotations for every complete circuit of the terrella:

a magnetick needle, moved on the top of the earth or of a terrella or of the effused orbes, makes two complete rotations in one circuit of its centre, like some epicycle about its orbit. (Gilbert 1600, Book 5, chap. 11)

Given this, then, magnetic declination is not a movement of deviation from the horizontal but rather “is in reality a motion of rotation” (Gilbert 1600, Book 5, chap. 6). The conclusion that declination is a rotation is not extracted out of one particular observation. It is rather prompted by a series of observations when (and only when) taken together.

I do not want to claim that Gilbert’s conclusion that declination is “in reality a motion of rotation” is fully constrained or determined by the experimental practice. Rather, given the investigative practice of mapping the phenomenon of declination, the claim that rotation is shown in declination is a cogent reading off of the experimental situation.36 If Gilbert had stopped at measuring the magnetic declination in a single location, for instance, it would have been quite difficult to conclude that the movement of declination is part of a rotation. The practice of mapping out declination around the terrella, however, made this conception likely. The experimental setup itself, the relative inadequacy of the instruments and an abiding assumption of order, all conjoined, suggested that, in orbiting the terrella, the needle rotates.

The conclusion from the investigation of declination, then, is that if one magnetic body orbits another, it will rotate around an axis as well. This serves to suggest that an orbiting Earth is a diurnally rotating Earth. However, the Earth’s axial rotation and its magnetic axis coincide (for Gilbert), but magnets, as shown in Gilbert’s investigation of declination, rotate around axes perpendicular to their magnetic axes.37 Because of this (alongside the general omission of the declination experiments and their relevance for magnetic rotation in the literature), some of Gilbert’s readers, such as Henry (2001), or Miller (2014) who calls Gilbert’s argument a “blind alley” (Miller 2014, p. 79), take it that Gilbert has no satisfactory argument for terrestrial axial rotation as a specifically magnetic rotation.38 There are two reasons for the dissatisfaction.

First, there is an assumption that rotation has to be not only continuous, but also spontaneous. Or, put differently, it is assumed that the kind of rotation under discussion is a self-generating rotation of a magnetic body taken in isolation. This is not Gilbert’s claim. First, continuous rotation of a single magnetic body in isolation is not possible under his account of magnetism as disponent (because all motion is only in relation to alignment with the surrounding bodies). Second, Gilbert explicitly rejects the possibility of any magnetic perpetual motion machine (Gilbert 1600, p. 107). Continuous full axial rotation requires the rotating body to be orbiting another celestial body. Rotation thus presupposes relationality, and, additionally, it is specific to the precise order of nature, an order from which celestial bodies cannot deviate (too much) without disrupting and collapsing the universe (Gilbert 1600, pp. 220–25; and Miller 2014, pp. 80–6).

Second, the dissatisfaction also comes from taking a strong Copernican reading of De magnete: the goal is to show that magnetism provides a complete “model” of terrestrial axial rotation. If that were Gilbert’s aim, he would need to show how magnetism causes rotation around the north–south magnetic axis. And he fails to show this. But this is an unnecessarily strong reading. I think Gilbert’s claims should be taken modally: magnetism might explain the Earth’s rotation because magnets can (given the appropriate conditions) rotate around an axis. Under this interpretation, De magnete does indeed show that magnetism can in principle produce the kinds of movements we see in celestial bodies, but it does not provide a full account of how magnetism causes the Earth’s axial rotation.

Moreover, it is also equally possible that magnetism as disponent is not causally sufficient to bring about full sustained rotation around the axis; something else might be necessary, as is the case for magnetic coition and its requirement for the “double vigour” of disposition and coition (see Section 5.4). This interpretation fits with Gilbert’s overall project: if one does not have an account of the circumstances that bring about a given phenomenon (the north–south axial rotation of the Earth, for instance), one does not yet have an explanation. Whilst the disponent power is necessary for the explanation, it might very well not be sufficient. This is not a problem for Gilbert if all he intends is to show the plausibility of axial rotation. The project for a fully-worked-out Gilbertian magnetic cosmology might be precisely the identification of the exact circumstantial conditions that would explain continuous north–south axial rotation.

Gilbert is aware of this; he takes his magnetic observations to provide new support for the Copernican model, but not to prove beyond doubt the “probable assertion” (Lat.: probabilis assertio) of the Earth’s diurnal rotation (Gilbert 1600, p. 215). The final book of De magnete, in which Gilbert tackles the plausible extension of his magnetic philosophy to Copernicanism, is more concerned with showing the plausibility of the daily axial rotation against primum mobile accounts than it is with showing how magnetism is supposed to fully explain the rotation. Additionally, it should not be forgotten that Gilbert’s commitment is to establish the premises of his arguments by experiment, and, in the introduction to the treatise, Gilbert mentions that on cosmological issues he allows himself “to philosophise freely and with the same liberty which the Egyptians, Greeks and Latins formerly used in publishing their dogmas” (Gilbert 1600, p. iii).39

Taken as a full physical explanation of the Earth’s diurnal rotation, then, magnetic rotation fails. But as an assertion of the plausibility of magnetism causing effects of the same kind as those witnessed celestially, it works. And the rotation Gilbert sees in the declination experiments is the best indication that this is so.

5.3. The Incompatibility between Mapping and the Attractional Framework

If Gilbert had understood magnetism as purely attractional, the mapping of magnetic declination around a terrella would not have been an available means of investigation. In fact, none of the previous models would have facilitated or favored an investigative practice with this goal. The attractional model, with its posited sources of magnetism at various locations either in the Earth or in the heavens, could not have accommodated the phenomenon of declination, as Norman had already suggested experimentally.40 Norman’s alternative conception, the respective point, did open up the investigative path of taking measurements at various geographical locations, with the intention of triangulating the position of the respective point. But Norman’s respective point is a source of magnetism that is specifically “a point along a line that passes through the Centre of the Earth” (my emphasis)41, and thus a laboratory-based mapping of declination around a loadstone would not have been a viable investigative path for studying the effect. For Norman, dip (his term for declination) was a specifically (exo)geomagnetic phenomenon; it had nothing to do with loadstones. For Gilbert, on the other hand, any two magnetic bodies could show declination, given the appropriate relative arrangement.

5.4. Magnetism as Disponent Affords Mapping Practices

I have shown that Gilbert took rotation to be a magnetic motion, fully manifest in the investigation of declination through the mapping experiments. The question is why the mapping experiments are a viable path of investigation. One answer could be that the use of a terrella as a model of the Earth is itself afforded by Gilbert’s matter theory (Section 1). While this explains the use of the terrella, it does not address why Gilbert takes mapping to be the appropriate investigative path. My position is that the mapping of magnetic phenomena in laboratory investigations becomes appropriate because Gilbert treats magnetism as primarily disponent, i.e., as ordering magnetic bodies. That is, there is an order to be mapped, and this order is fundamentally revealing of magnetic power. This treatment of magnetism as disponent, I argue here, was afforded by Gilbert’s studies of magnetic attraction.

Whilst studying magnetic coition, Gilbert makes a conceptual distinction between (a) the power of coition (coitio vigor/vis) and (b) the disponent power. For the phenomenon of coition to occur, the magnetic bodies must first align; once aligned, if within the orb of coition, and not just within the orb of virtue, they unite. One can reasonably ask if there is anything empirical that prompted this distinction between the two. My best reconstruction is that the distinction resulted out of Gilbert’s investigations of magnetic repulsion.

When treating repulsion, Gilbert concludes that, “magnetick substances are more sluggishly repelled than they are attracted” (Gilbert 1600, p. 100). This is not the result of a specific experiment, but a generalization over several experiments: “[it] is manifest in all magnetical experiments in the case of stones floating on water in suitable skiffs; also in the case of iron wires or rodes swimming (transfixed through corks) and well excited by a loadstone, and in the case of versoria” (Gilbert 1600, p. 100). In other words, irrespective of the experimental setup, the repulsion is always slower than the coition of the magnetic bodies. This is an empirical regularity that needs to be accounted for. He accounts for it thus: “repulsion and aversion is caused merely by something disponent; on the other hand, the coming together is by a mutual alluring to contact and a disponent, that is, by a double vigour” (Gilbert 1600, p. 100; my emphasis). What we observe as repulsion is for Gilbert a mere arrangement of magnetic bodies—that is, an aligning of their poles—so as to be able to unite.

In a strict sense, then, for Gilbert, magnets do not repel (as we would have it today): what looks like repulsion is arrangement so that coition is possible. This is not such a forced interpretation on Gilbert’s part: after all, if you place two magnets at appropriate distances they will eventually unite, even if they appear to repel at first. Repulsion in itself is never a stable event, but it always leads to the opposing poles’ coming together (unless prevented by distance or external forces). And this is the reason behind Gilbert’s explaining away repulsion as a magnetic phenomenon, and merely treating as the effect of the disponent power, which arranges, or organizes, the magnetic bodies before their coming together.

Gilbert breaks down an observational situation previously taken to be a single event, the attraction between magnetic bodies, into two distinct events: one is the coming together of magnetic bodies,42 and the other is the aligning (disposing) of magnetic bodies in such a way that their coming together is possible (the north–south alignment). He thus takes the alignment to be an independent event to be accounted for. The “mutual alluring to contact” (the power of coition) is what explains the empirical law of coition: that the north pole always attracts the south pole, and vice versa, and can only take place when magnetic bodies are placed within the “orb of coition.” The disponent power, on the other hand, explains how magnetic bodies organize themselves so as to eventuate the unition of magnetic bodies (if conditions meet). There are then (at least) two possible manifestations (expressions) of the magnetic power: as disposing magnetic bodies and as bringing together magnetic bodies—as disponens vigor and as coitio vigor.

As we have seen in the case of experimental observations of the sluggishness of repulsion, the disponent power is a precondition to coition. First, magnetic bodies arrange (in their proper relative positions), and then they unite so that “a disponent vigour [power] is often only the precursor of coition, in order that the bodies may stand conveniently for one another before conjunction; wherefore also they are turned around to the corresponding ends, if they cannot reach them through hindrances” (Gilbert 1600, p. 101).

The sluggishness of magnetic repulsion is doubly important for Gilbert: not only because it gives an impetus to explain magnetic repulsion away, but also because it allows him to show that the disponent power is truly a basic manifestation of the magnetic power. Had Gilbert concluded that attraction and repulsion are exactly equal, then he probably would have had more pressure to account for both events through coition.43 But his experimental identification of the sluggishness of repulsion led him to differentiate a disponent power as the manifestation of the magnetic power, in the context of the laboratory study of magnetic attraction.

5.5. Full Circle: from Attraction to Rotation

As we have seen, the relation between Gilbert’s investigative practices and his treatment of magnetism as disponent is as follows. Following the order of presentation of the arguments in De magnete, one first begins with a pattern of investigation that is congruent with the attraction framework of studying magnetism (the experiments involving magnets floating on water), which boils down to tinkering with magnetic attraction (and repulsion) to establish claims about these effects, or about what they show. In this context the disponent power is introduced as a result of a phenomenal differentiation between the movement to unition of two magnetic bodies, and a quite distinct movement to disposition. This conceptualization of magnetic phenomena as relative alignments of magnetic bodies guides the local investigation and the further theorization of magnetism—one that supplants the attractional model (Section 3).

By treating magnetism as disponent, new directions for investigating magnetism are opened up in the form of an added epistemic goal: to map out the spatial arrangements of observed magnetic phenomena (Section 5.3). Heuristics aimed at spatial mapping take center stage: to account for declination is to give a model of the spatial relations it forms, and then to draw regularities out of this model (in the form of a nomograph that calculates the declination–latitude relation). What is taken to be appropriate or inappropriate as an investigative path is judged against a conception of magnetic phenomena as primarily phenomena of relative arrangements.

In De magnete, the mapping heuristics are substantiated through versorium-based experiments. The versorium detects the motion, while specific apparatuses produce the appropriate conditions for the detection to be readable. It is in this context, and through the investigation aimed at mapping out the magnetic motions of declination, that the phenomenon of magnetic rotation is afforded. The specification of magnetic rotation presupposed conceptualizing magnetism as disponent. Moreover, it is through the mapping experiments that magnetic rotation is made apparent (Section 5.3). But Gilbert would have no mapping experiments without treating magnetism as alignment (Section 5.4); and the latter was facilitated by the floating magnet experiments that show a disponent power separate from and prior to the power of coition (Section 5.4). If one does not accept the versorium-based mapping practice of investigating magnetic declination (as well as direction and variation) as legitimately detecting the magnetic phenomena then one does not accept the claim about magnetic rotation.

6. Conclusion

I have shown that Gilbert moves away from the conception of magnetism as fundamentally attractional to a conception of magnetism as disponent. The latter accounts for the observed magnetic phenomena. (In the case of magnetic coition, the disponent power is a co-cause, and takes temporal priority.) I have also shown that magnetism as disponent was a conceptual move afforded by Gilbert’s treatment of magnetic action as symmetrical, with every magnetic body of an interaction contributing. The system of magnetic bodies takes priority over the individual bodies. Once we shift attention to magnetism as disponent, we can make better sense of Gilbert’s claim that magnets rotate and that, consequently, the Earth’s diurnal rotation is due to its magnetic nature (however that may work in detail). This is because the disponent model makes the mapping of magnetic effects a suitable means of investigation. And it is in the mapping of declination around a terrella that Gilbert finds an experimental manifestation of complete, continuous magnetic rotation. It is ultimately through understanding Gilbert’s conception of magnetism as fundamentally disponent that we can see the warrant for magnetic terrestrial rotation.

Notes

1. 

Disponens is the present participle of the verb dispōnō. Dispōnō means “to dispose,” “to place here and there,” “to arrange,” “to regulate,” “to distribute,” “to administer/manage/order.” Despite the various ways in which the verb can be translated, all of these meanings seem to have the following in common: they seem to refer to situations of arranging for action, or into place, or following into a system of order.

2. 

Gilbert’s followers systematized Gilbert’s magnetic philosophy, and in the process continuously transformed Gilbertian concepts. This is not too surprising: Gilbert’s text was far from being consistent enough not to allow for different interpretations or re-workings of the claims made in De magnete. Re-workings of arguments often lead to transformations of those arguments. This is precisely what both Ridley (1613), Carpenter (1635), and others did. In the attempt to clarify Gilbert’s theory, they changed Gilbert’s theory. These changes are not however the subject of this paper. For my purposes here what matters is that readings of Gilbert’s theories as disponent were also available to Gilbert’s followers; that is, I am not alone in bringing attention to such a conceptualization of magnetism.

3. 

On Gilbert’s Copernicanism see Bennett 1981; Henry 2001; Miller 2014. On responses to the magnetic cosmology see Baldwin 1985.

4. 

On the discipline of magnetic philosophy, see Pumfrey 1987.

5. 

On Gilbert’s matter theory, see Roller 1959 and Freudenthal 1983.

6. 

Gilbert uses the versorium to investigate the magnetic motions. Unmagnetized, the versorium is used as a crude electroscope to detect the static electricity of various substances and metals in chap. 2, Book 2 of De magnete.

7. 

A tilted magnetic dipole takes the magnetic axis to be tilted relative to the Earth’s axis of rotation, while an axial magnetic dipole conflates the magnetic axis with the rotational axis. For various accounts of axial and tilted dipoles from the sixteenth century to the nineteenth century, see Jonkers 2003, chap. 2.

8. 

“New” here is not to be read absolutely. It refers to the construction of a new word, but also to the appropriation of a word from a different domain and tailoring it for magnetism. This latter strategy is very frequently put to use in De magnete. Vis disponens is not an absolutely new concept; it is reappropriated from, most likely, medieval theories of the soul, as Nathanael Carpenter argues (1635, p. 48). I would like to thank an anonymous referee for raising this important point. The intellectual origins of this concept would be a worthwhile contribution to the discussion, but it goes beyond the extent and interests of this paper.

9. 

As part of the perceptual disclosing, De magnete makes use of diagrams. See Georgescu 2014.

10. 

The actual story is a little more complicated than this summary might imply, but the overall point holds.

11. 

Unwarranted distinctions at this level are prone to error. Whether something is unwarranted or not is a normative question, and it can only be set by norms of experimental practice or norms for what is taken to be a reliable knowledge claim at any given point. For an excellent treatment of how the spatial description of a magnetic body guides Gilbert’s inquiries see Miller 2014, chapter 3.

12. 

Aristotle’s philosophy of change and its constant reworking, especially in medieval philosophy, is a large topic, and is treated only superficially in De magnete. Its subtleties, however, are not of concern here.

13. 

On pre-Gilbert theories of magnetism, see Smith 1968, 1992, and Jonkers 2003, especially chap. 2.

14. 

Although frequently invoked in studies of the long-term history of magnetism, Norman’s short treatise has not yet received the attention it deserves. Harré (1981) is an exception and offers a nicely detailed examination of Norman’s experiments.

15. 

De mundo is a thorough step-by-step rejection of Aristotelian cosmology and physics. There is currently no English translation of De mundo, although a translation by Stephen Pumfrey and Ian Stewart is in preparation. For a detailed study of De mundo, see Kelly 1965.

16. 

My translation. The Latin reads: “Ad locus non operatur, nec dominatur in rerum natura, ut sistat corpora, aut moveat. […] [L]ocus enim nec ens est, nec efficiens causa: ab ambientibus corporibus saepe vis influit in contentum.”

17. 

I know of no thorough treatment of Gilbert’s theory of motion. On his conception of place, see the cursory remarks of Cassirer (2003, p. 362).

18. 

Gilbert seems to believe that the concept of sympathy does not capture his conception of mutuality because (1) sympathy is not explanatory and (2) the logic of sympathy makes sense only when conceptualized alongside antipathy, and the latter is not a proper action of magnets.

19. 

Or, in the case of coition, initially through the disponent power.

20. 

For treatments of Gilbert’s magnetic power see King 1959; Hesse 1960; Freudenthal 1983; Pumfrey 2002. None of these treatments are exhaustive or conclusive.

21. 

This suggestion is not made mainly on the basis of precise textual evidence (since De magnete rarely provides what we would traditionally qualify as philosophical arguments), but rather on the basis of how we can consistently make sense of how these concepts are put to work.

22. 

Reading it in this way makes sense of why Gilbert uses the “vigor,” “vis,” and “potentia” in multiple different terms. He can be more liberal with how he terminologically describes the magnetic power in its manifestation.

23. 

Neither Thompson nor De Mottelay, the two English translators of De magnete, is consistent in translating the verb dispōnō and its correlates. I have chosen to stick to Thompson’s translation throughout the text, as it is closer to the original text.

24. 

The treatment I offer might also explain why Gilbert’s talk about the magnetic power is metaphysical (in terms of intelligences or astral forms), while the investigation of the magnetic power’s manifestation as disponent is in terms of physical entities (meridians, poles, sizes, relative distances, orbs, etc).

25. 

Notice also that what a surrounding body is, is expressible relative to the combined strengths of the magnetic bodies (i.e., the extent of their respective orbs of virtue).

26. 

I suggest that the Latin, vigor/virtus disponens, supports my reading. See fn.1.

27. 

Gilbert’s animism has been a point of focus in the literature: it was a marker of his “pre-scientificism” (King, 1959), an “emotional background” that does not affect the empirical content of the treatise (Zilsel, 1941, p.6), while Henry (2001) argues for complementarity between the experimentalism and what he calls the “critical” animism of the treatise. It is no easy task to settle the debate on Gilbert’s animistic doctrine, especially when the arguments are used as subservient to other goals in the scholarship: for instance, showing that De magnete belongs to the natural magic tradition in Henry’s case, or showing that the craftsman played an essential role in the advent of modern science, in Zilsel’s case.

28. 

Gilbert took the concept of orb from astronomy and applied it to magnetism. But then, he reverses the relation: an orb is a concept pertaining to a magnetism-based physics that is extended to astronomy because, in astronomy, we observe and describe the manifestation of magnetism on the cosmological scale. It should be noted that the orb is only a subsistent (its existence depends on the interaction of magnetic bodies as seen in chap. 11 of Book 5).

29. 

It might seem as though Gilbert’s notion of verticity is more relevant than that of the disponent power here, given that he explicitly states that magnetic direction is caused by verticity (Gilbert 1600, p. 119). I take it that verticity is secondary to the disponent power, at least in this case. Verticity is what gives polarity to a magnet; it orients the magnet in a north–south position (Gilbert 1600, pp. 119–22). This makes it seem like a directive, aligning force. And, to an extent, it is. But, on Gilbert’s account, poles (in magnetic bodies) are established and can change only relative to interaction within a system of magnetic bodies; a magnetic body gets its poles with respect to its alignment with another magnetic body (this is true of the magnetization of iron as well as of loadstones in their initial locations in the earth). So, while verticity indeed describes the magnetic force in respect of producing poles, its manifestation is still ultimately one of disponens, as it directs and organizes. In addition, the north–south orientation is only specifiable relative to a system of bodies (this applies to all magnets including the Earth, although, admittedly in the Earth’s case this orientation in the universe is unchanged). Gilbert also makes explicit reference to magnets disposing themselves in the case of verticity (Gilbert 1600, pp. 124, 125, 127, 131, etc.).

30. 

This is a direct consequence of his extrapolation of the distribution of magnetism on the terrella to the Earth itself.

31. 

The assumption that magnetic bodies align according to an order is a consequence of an overall teleology of the law of the whole that ultimately governs Gilbert’s version of a system of the world. I will not develop this point further, but in my interpretation the order of the universe acts as a final cause for local orders.

32. 

So far, the literature has not much discussed the extent to which teleology and the system of the world play a role in Gilbert’s cosmology specifically, and to his overall arguments more generally. I take Miller’s discussion of the law of the whole to be a notable exception (Miller 2014, pp. 81–6).

33. 

Gilbert mentions some experimental setups in Book 5, all of which are directed towards the mapping of dip. For instance, Gilbert details an experimental setup in which a magnetized needle is suspended in air, with a marked terrella underneath; the terrella is moved in various positions relative to the N–S axis of the poles, and the deviation of the needle is observed.

34. 

In fact, in Gilbert’s experimental setup, the terrella is placed in a bowl-shaped hollow, which allows it to be rotated in place while the versorium stays put (see Georgescu 2014 for more on the setup itself). This makes for simpler, more reliable experimental manipulation, but the effect is that, relative to the terrella, the versorium changes latitude.

35. 

This is Ridley’s formulation of the declination-latitude correlation (1617, p. 21).

36. 

I use a conceptual distinction between investigative and laboratory/experimental practices. Laboratory or experimental practices refer to the material interventions, whereas the investigative practices include the laboratory practices, but also the various procedures and techniques in one’s repertoire for translating the laboratory findings into results; that is, the diagrammatic, analogical, etc. reasoning that informs the reading off of the result from the experimental situation. The distinction is artificial: there is no recorded experimental situation that is not already filtered by background knowledge and techniques of translation since any recorded experimental result is a retrospective and idealized account.

37. 

The axial and magnetic rotations coincide because of the geometry of the magnetic power’s distribution on the surface of the terrella by appeal to magnetic poles, a magnetic equator, and magnetic meridians. On these points see (Gilbert 1600, Book 2); and Miller 2014.

38. 

The debate on these issues seems to me to come down to what one takes a satisfactory argument to be. Gilbert does not seem to operate with clear-cut criteria. In this paper, I have therefore proposed a reconstruction, which attempts to make explicit what sort of findings might have presented themselves as showing magnetic axial rotation, or at least as rendering it plausible. It is by no means a final answer to the debate.

39. 

The claim is not that Gilbert is not committed to the diurnal motion of the Earth. Book 6 makes it clear that Gilbert takes Copernicus’s astronomical model to be more likely than the Primum Mobile model. He also considers that his magnetic philosophy could contribute to explaining why this is the case, since the Earth’s magnetism shows it to be “sufficiently furnished with peculiar forces for diurnal circular motion” (Gilbert 1600, p. 223). The claim here is just that Gilbert does not fully subordinate his magnetic theories to showing that the diurnal rotation of the Earth is caused by magnetic rotation. Book 6 of De magnete does not discuss how or why magnetic rotation accounts for the Earth’s diurnal rotation; it only overviews reasons (some drawn out of his magnetic philosophy, some not) for why he takes the Earth to rotate axially.

40. 

Norman’s experimental setup is devised to verify the behavior of the needle as predicted within the attractional framework, and the setup is adjusted to this end (see Norman, 1581, pp. 21–2).

41. 

Norman 1581, p. 28.

42. 

The first instance we know of in which the law of attraction is formulated is in Peter Peregrinus’s Letter on the magnet ([1269] 1904). For treatments of Peregrinus see Radelet-de-Grave and Speiser 1975 and Steinle 2012.

43. 

How much epistemic warrant there is for Gilbert’s reading of the sluggishness of repulsion is a different question, whose answer is independent of the larger point I am trying to make here.

References

Baldwin
,
Martha
.
1985
. “
Magnetism and the Anti-Copernican Polemic
.”
Journal for the History of Astronomy
16
:
155
74
.
Bennett
,
Jim A.
1981
. “
Cosmology and the Magnetical Philosophy
.”
Journal for the History of Astronomy
12
:
165
77
.
Carpenter
,
Nathaneal
.
1635
.
Geographie Delineated Forth in Two Bookes
.
Oxford
:
Henry Cripps
.
Cassirer
,
Ernst
.
2003. (republished)
Substance and Function & Einstein’s Theory of Relativity
.
New York
:
Dover Publications
.
Freudenthal
,
Gad
.
1983
. “
Theory of Matter and Cosmology in William Gilbert’s De magnete
.”
Isis
74
:
22
37
.
Gaukroger
,
Stephen
.
2006
.
The Emergence of a Scientific Culture: Science and the Shaping of Modernity, 1210–1685
.
Oxford
:
Oxford University Press
.
Georgescu
,
Laura
.
2014
. “
The Diagrammatic Dimension of William Gilbert’s De magnete
.”
Studies in History and Philosophy of Science
(
Part A
)
47
:
18
25
.
Gilbert
,
William
.
1600
.
De magnete, magnetisque corporibus, et de magno magnete tellure; Physiologia nova, plurimis & argumentis, & experimentis demonstrata
.
London
:
Peter Short
.
Gilbert
,
William
.
1651
.
De Mundo Nostro Sublunari Philosophia Nova
.
Amstelodami
:
Ludovicum Elzevirium
.
Harré
,
Rom
.
1981
.
Great Scientific Experiments: 20 Experiments that Changed Our View of the World
.
Oxford
:
Phaidon
.
Henry
,
John
.
2001
. “
Animism and Empiricism: Copernican Physics and the Origin of William Gilbert’s Experimental Method
.”
Journal of the History of Ideas
62
:
99
119
.
Hesse
,
Mary B.
1960
. “
Gilbert and the Historians I & II
.”
British Journal for the Philosophy of Science
11
:
1
10
;
130–42
.
Jonkers
,
A.R.T.
2003
.
Earth’s Magnetism in the Age of Sail
.
Baltimore and London
:
The Johns Hopkins University Press
.
Kelly
,
Sister Suzanne
.
1965
.
The De Mundo of William Gilbert
.
Amsterdam
:
Menno Hertzberger and Company
.
King
,
W. James
.
1959
. “
The Natural Philosophy of William Gilbert and his Predecessors
.”
United States National Museum Bulletin
218
:
121
39
.
Michell
,
John
.
1750
.
A Treatise of Artificial Magnets
.
Cambridge
:
J. Bentham
.
Miller
,
David Marshall
.
2014
.
Representing Space in the Scientific Revolution
.
Cambridge
:
Cambridge University Press
.
Norman
,
Robert
.
1581
.
The Newe Attractive, Contayning a short discourse of the Magnes or Loadstone
.
London
:
Richard Ballard
.
Peregrinus
,
Petrus
.
[1269] 1904
.
The Letter of Petrus Peregrinus: On the Magnet
, trans. by
Br.
Arnold
,
New York
:
McGraw Publishing Company
.
Pumfrey
,
Stephen
.
1987
.
William Gilbert’s Magnetic Philosophy: the Creation and Dissolution of a Discipline
.
University of London: Unpublished PhD Dissertation
.
Pumfrey
,
Stephen
.
2002
.
Latitude and the magnetic earth
.
Duxford, Cambridge
:
Icon Books
.
Radelet-de-Grave
,
Patricia
and
Speiser
,
David
.
1975
. “
Le De Magnete de Pierre de Maricourt. Traduction et Commentaire
.”
Revue d’histoire des sciences
28
(
3
):
193
234
.
Ridley
,
Mark
.
1613
.
A Short Treatise of Magneticall Bodies and Motions
.
London
:
Nicholas Okes
.
Ridley
,
Mark
.
1617
.
Magneticall Animadversions
.
London
:
Nicholas Okes
.
Roller Duane
,
H.D.
1959
.
The De magnete of William Gilbert
.
Amsterdam
:
Hertzberger
.
Smith
,
Peter J.
1968
. “
Pre-Gilbertian conceptions of Terrestrial Magnetism
.”
Tectonophysics
6
(
6
):
499
510
.
Smith
,
Julian A.
1992
. “
Precursors to Peregrinus: The Early History of Magnetism and the Mariner’s Compass in Europe
.”
Journal of Medieval History
18
:
21
74
.
Steinle
,
Friedrich
.
2012
.
Goals and Fates of Concepts: The Case of Magnetic Poles
. Pp.
105
126
in
Scientific Concepts and Investigative Practices
. Edited by
Uljana
Feest
and
Friedrich
Steinle
.
Berlin
:
De Gruyter
.
Zilsel
,
Edgar
.
1941
. “
The Origins of William Gilbert’s Experimental Method
.”
Journal of History of Ideas
2
:
1
32
.

Author notes

Research for this paper was funded by grant 11K8913N from the Fonds Wetenschappelijk Onderzoek—Vlaanderen. My thanks to Eric Schliesser, Katherine Brading, Monica Solomon and Barnaby Hutchins for comments on earlier drafts. I am especially grateful to Stephen Pumfrey and an anonymous referee for their thorough and insightful comments.