This Newsletter was published on Saturday 3 July 2010

The Cartography of “The Definition of Love” Revisited

The “Definition of Love” is one of the most highly regarded and widely disseminated of Marvell’s poems. Critics concur in its praise, anthologists in reprinting it. Its appearance in all but one of the standard student anthologies of the period (the Norton, the Longman, the Oxford, but not the Broadview) suggests that it is often taught, and one might presume from this that there is consensus over its method and meaning. A review of the commentary tradition shows otherwise.

While commentators are unified in their praise of the poem, there are disputes over its genre–in what sense might it be, as its title insists, a definition?–and over its mode of argument. Critics agree that its excellence is related to its integration of abstraction and concretion, but they disagree over whether the poem turns abstract ideas into concrete details, or the other way around (Leishman 68; Schmitter and Legouis 50 n. 1, 53; Legouis, Andrew Marvell 76). A solution to this dispute can only follow from a clear sense of how the central conceits of the poem hang together: on how the poem leads its reader from the “strange and high” object of stanza I, through the “extended soul” and fate’s “iron wedges” of stanza II, the “distant poles” of stanza V, the planisphere of stanza VI, and the lines oblique and parallel of stanza VII, to the mental conjunction and sidereal opposition of the concluding stanza (Marvell 109-11). To Leishman such conceits form no chain of reasoning or argument whatever: they are just several witty ways of saying “‘We can never meet'” (69). Others have tried to establish an internal logic among them. The seminal account was furnished (as often) by Legouis, who in his 1928 dissertation noted a move from celestial cartography in stanzas V and VI to geometry in stanza VII, ending with astrology in stanza VIII (André Marvell 145-6). But a subsequent suggestion that the poles and planisphere of stanzas V and VI were terrestrial rather than, as Legouis had originally assumed, celestial, led him to revise his position and to add, in the 1965 English abridgement of his work, terrestrial cartography as a bridge between the astronomy of stanza V and the geometry of stanza VII (Andrew Marvell 75 n. 1). (He did not explain the implications of this additional point of reference for his claim about the poem’s striking unity.)

Legouis’s concession on this point has been followed by most subsequent critics and editors. But it is worth asking whether the arguments in favor of a terrestrial reading are sufficiently compelling to warrant its status as a serious alternative to the celestial. A reexamination of the critical history suggests that scholars–including Legouis himself–have been too quick to embrace it.

Dennis Davison was, I think, the first to build a substantial argument for the coherence of stanzas V-VIII based upon the assumption that the poles of stanza V are celestial. Davison claimed in 1955 to have demonstrated a hitherto unsuspected unity in the poem by showing that stanzas V-VIII are exclusively astronomical in scope (in so doing he was apparently ignorant of Legouis’s priority in advancing such a claim) (“Marvell’s ‘The Definition of Love'” 143). Davison suggested that the distant poles of stanza V are the celestial poles, that the world which wheels about them is the sphere of the heavens, that the planisphere of stanza VI continues the celestial reference, that the oblique and right lines and angles of stanza VII refer not to geometry but to astronomy and astrology–such terms having specific meanings in those contexts, and that the concluding stanza, with its reference to opposition and conjunction, is also astronomical and astrological in reference. Davison asserted that, read in astrological terms, “the final four stanzas form a logical sequence of related images”; but he was uncertain what logic such a sequence might entail (145).

Davison was sharply rebutted half-a-dozen years later by Dean Morgan Schmitter, who insisted in a note on the cartography of the poem that the imagery of stanzas V-VII was terrestrial. He suggested that the oblique and parallel lines of stanza VII could well be terrestrial on the grounds that a planispheric map of the earth would show lines of latitude running in parallel circles and lines of longitude running obliquely and forming angles where they cross at the poles (Schmitter and Legouis 49). And he offered at least three reasons for preferring a terrestrial reading. First, “the idea of common loves would be embarrassed by a reading that links them with cosmic values.” Second, the terrestrial version of the conceit is more familiar and easier to visualize than the celestial, “which can only be imagined as projections of the earth’s axis into a constantly changing space” (50); terrestrial poles, by contrast, are easily imaginable, and the flattening of the globe to create a double-faced surface–one centered on the north and the other on the south pole, and each showing the northern or southern hemisphere in stereographic projection–would have been familiar to Marvell’s first readers from the popular atlases of the age (49-51). And third, the term “world” (line 19) more commonly refers to the earth than to the cosmos as a whole (51).

Schmitter received reluctant support in an accompanying note by Legouis, who conceded that most of the arguments for a celestial reading could as well support a terrestrial one and accepted (after failing to find any celestial maps featuring both obliques and parallels) Schmitter’s suggestion that the oblique and parallel lines were the meridians and parallels of a terrestrial map as a means of connecting stanza VII to stanza VI, with its concluding reference to a planisphere (Schmitter and Legouis 52). But Legouis was unwilling to accept Schmitter’s claim that the terrestrial version would have been more familiar to contemporaries than the celestial because, he suggested, stereographically projected maps of the heavens were probably more common than those of the earth. In a long footnote he recorded his impression, derived from forays into the map rooms of several archives, that most terrestrial atlases of the age did not show the northern and southern hemispheres in full projection and that those maps which did use such projections were usually celestial (52-3 n. 3).

Although Legouis clearly wanted to resolve the matter in favor, as far as possible, of his own earlier “celestial” reading of the poles, he was distracted and put out by Davison’s failure to acknowledge his priority in advancing such a reading (51 n. 1), unimpressed by what he termed Davison’s “far-fetched and rather confused interpretation” of stanza VII on such lines (52 n. 1), and unable definitively to counter Schmitter’s evidence. He therefore left the matter scrupulously in the balance, both in his 1961 note, and the notes to his influential 1971 revision of Margoliouth’s Oxford English Texts edition of Marvell (1:259, l. 18). Subsequent scholarship has generally and wisely followed him. Even Davison, in his 1964 student guide to Marvell, did so (Poetry 34).

It is unfortunate that Legouis didn’t spend a little longer investigating the matter, because his reservations were justified: Schmitter’s three arguments for reading stanzas V-VII as terrestrial are not compelling enough to warrant rejecting the prior and more obvious celestial interpretation. First, his claim that the “the idea of common loves would be embarrassed by a reading that links them with cosmic values” derives from a common tendency (recently criticized by George Klawitter) to read the poem through the lens of Donne’s “A Valediction: forbidding Mourning,” with its distinction between celestial and “[d]ull sublunary lovers’ love” (Donne 84). Marvell, unlike Donne, draws no distinction between celestial and sublunary lovers; he distinguishes, rather, those whose lines are parallel from those whose are oblique. The logic of Schmitter’s claim is self-cancelling: were these parallel lovers placed at the north and south terrestrial poles they would be no less terrestrial than any others.

Second, we should have reservations about accepting Schmitter’s terrestrial explanation for the planisphere of stanza VI, to which three objections may be raised. The first two are technical. First, as Schmitter himself notes, polar stereographic projection would show every line of latitude as a perfect circle in parallel to every other line of latitude (49). On such a map, there would be nothing distinctive about a circular line, and this would fail to emphasize the distinctiveness of the lovers. Second, the oblique longitudinal lines would “meet” one another only at the poles, and not, as the text insists, “in every angle.” The third objection is historical. While it is true that Marvell’s contemporaries were familiar with stereographic projection in terrestrial cartography, it is wrong to suggest that they were unfamiliar with polar stereographic projections in celestial cartography. Polar stereographic projection was–as Legouis suspected–the norm in celestial cartography (Snyder 22-3). In terrestrial cartography the standard projection was not polar but equatorial–most of the well-known world maps of the age (those of Mercator, Speed, Blaeu, and Hondius) being so projected (Snyder 27). One contemporary world map to adopt a polar projection was that of Cornelis de Jode, appearing in his Speculum Orbis Terrae. A glance at it shows why this was an unusual point of view: at the center of the Antarctic region is a large blank continent entitled “Ter. Australis incognita”; at the center of the Arctic region, at the intersection of an empty, quadripartite continent is a large rock, then thought to be the north pole. De Jode’s atlas appeared in 1593 and was never reprinted (Baynton-Williams 49). Another instance of polar projection appears in Jan Jansson’s Atlas Novus, which fits small polar projections of the northern and southern hemispheres in the gaps between its eastern and western hemispheric projections: the polar projections are, in other words, subordinate to the main maps and are primarily decorative (Nordenskiöld). Those who favor a terrestrial reading of the planisphere must furnish stronger evidence than Schmitter was able to muster for the claim that Marvell’s contemporaries would have been more familiar with the earthly globe than with the cosmos stereographically represented in polar projection.

Finally, Schmitter’s claim that the term “world” is more frequently employed for the earth than for the universe as a whole seems to me neither here nor there as evidence for Marvell’s use of it in this poem. And a closer look at the definition Schmitter cites as evidence for his position raises doubts: “the earth and all created things upon it; the terraqueous globe and its inhabitants” (OED II. 7; quoted Schmitter and Legouis 51). While this definition, with its emphasis on the inhabitants of the earth as well as the globe itself, is not entirely unapt as way of reading the poem’s reference to the world’s being “cramped into a planisphere,” the cosmic definition, on the other hand, furnishes a better fit: “The material universe as an ordered system; the system of created things; ‘heaven and earth’; the cosmos” (OED II. 9). This definition, with its invocation of orderliness and systematic integration, squares closely with the poem’s emphasis on unprecedented disruption and disorder (“giddy heaven,” “new convulsion”). And its integration of “‘heaven and earth'” fuses the foci of the two preceding lines: “heaven” (line 21) and “earth” (line 22). External evidence helps buttress this claim: in his popular guide to the mathematical sciences, Thomas Blundeville uses the term “world” for the cosmos (136v).

But what of Legouis’s inability to pin down the celestial reading by way of celestial cartography? Legouis, as I mentioned, was unable to find contemporary star maps which showed latitudinal as well as longitudinal lines to allow for their meeting at angles as is required by stanza VII (“in every angle greet”). “But for this anomaly,” he noted, “I should not hesitate to assert that Marvell had celestial maps before his mind’s eye if he had anything concrete at all” (Schmitter and Legouis 53 n. 3). In searching for such lines on maps he was looking in the wrong place. As Margoliouth pointed out in a note to his 1927 edition of Marvell’s poems, “planisphere” was a common substantive term for an astrolabe–a device which, in the words of Blundeville, “is called of some a planispheare, because it is both flat and rounde, representing the Globe or Spheare, hauing both his Poles flatte both together” (Margoliouth 1:224; Blundeville 281r). Ann Berthoff picked up on this suggestion and posited a connection to the geographical astrolabe: an instrument in which a star chart is superimposed upon a map of earth (101), deploying an image of such a device for the dust-wrapper of her well-known book, The Resolved Soul. But the geographical astrolabe appears to be a much rarer instrument than the celestial version (only three of the 164 astrolabes listed in the inventory of the largest single astrolabe collection are geographical [Museum of the History of Science, Inventory nos. 44359, 53211, 53966]), and only the critical presumption that the poles are terrestrial obliges Berthoff to reach for it. As it happens, the standard western astrolabe, of the classic or planispheric variety, with a polar projection, contains all we need to overcome Legouis’s hesitation.

The planispheric astrolabe is projected stereographically from the imagined point of view of an observer at one of the celestial poles, looking directly up towards the opposite pole and the surrounding heavens, imagined as the inside of a cosmic sphere (Webster 29-30; National Maritime Museum 15). For obvious reasons, almost all astrolabes of this period are projected from the south celestial pole, thus showing the heavens in the northern hemisphere. The astrolabe itself is thus centered on the celestial poles–both of them, imagined as superimposed on one another from the point of view of an observer located at the southernmost celestial pole. This familiar device stands as strong evidence against the suggestion that Marvell’s contemporaries would have had difficulty locating or imagining the celestial poles.

The astrolabe also displays the mix of oblique and parallel lines Legouis sought and failed to find on contemporary star maps as a means of bridging between the planisphere of stanza VI and the geometry of stanza VII. The astrolabe is a device for determining star positions. To facilitate this, engraved on its circular plate or tympan is what looks like an oval net made up of oblique lines–oblique in the sense that they diverge from straightness (OED I. 1). This net divides up the visible sky in order to give celestial coordinates for an area defined by an observer’s horizon at a particular latitude. It shows circles of equal altitude (almucantars) and lines of direction (azimuths) which meet at the observer’s zenith. The angle of the earth means that the zenith is not the celestial pole, and the resulting distortion renders them oblique (National Maritime Museum 16-17; Webster 30, 34).

Not only do these lines create angles when they cross one another, the plates of many astrolabes were engraved with lines showing the divisions of the twelve Great Houses of the astrologers, the four cardinal houses of which were, as Davison notes, known as “angles” (OED n.2 5). In any given plate, three of these four angles will overlay the azimuths and almucantars. (The fourth, being below the horizon, falls outside the horizonally-defined net.) Such lines will thus greet one another in three out of the four “angles” (the first, seventh, and tenth houses) (National Maritime Museum 18; Webster 34).

In addition to presenting a distorted projection of the azimuths and almucantars, the tympan also shows, as a series of perfect concentric circles around the celestial poles, the Tropic of Cancer, the celestial equator, and the Tropic of Capricorn. These circles run parallel to one another and do not meet (National Maritime Museum 19; Webster 34). The astrolabe thus preserves the emphatic distinction between everyday obliquity and extraordinary parallelism implied by stanza VII and furnishes the circular parallels Legouis had failed to find on the star maps, thus overcoming his qualms about rejecting the terrestrial reading.

In the absence of additional evidence in favor of a terrestrial reading for the cartography of “The Definition of Love,” I suggest that we conclude from this brief review of the critical debate that the imagery in its latter stanzas is consistently celestial. Such a conclusion allows us to recognize that Davison was right in imagining that stanzas V-VIII do not shift their ground from celestial to terrestrial perspectives and back again, and that Legouis was right in sensing their “striking unity.” While this recognition might make life a little easier for future editors and glossators of Marvell by cutting out an unnecessary complication, such simplification is merely a prolegomenon to a renewed attempt on the fundamental critical problems, which remain to be satisfactorily solved: How, exactly, does the poem’s argument work? And in what sense is it a definition?

[Note: This article forms part of a paper presented at a meeting of the Andrew Marvell Society at the South-Central Renaissance Conference, Corpus Christi, TX, March 18-20, 2010. I am most grateful to my audience and co-presenters for their suggestions. I am also indebted to Ryan Mattke of the John R. Borchert Map Library at the University of Minnesota for cartographical assistance.]

Timothy Raylor
Carleton College

Works cited

  • Baynton-Williams, Ashley, and Miles. New Worlds: Maps from the Age of Discovery. London: Quercus, 2006. Print.
  • Berthoff, Ann E. The Resolved Soul: A Study of Marvell’s Major Poems. Princeton, N.J.: Princeton University Press, 1970. Print.
  • Blundeville, Thomas. His Exercises, containing sixe Treatises. London, 1594.
  • Davison, Dennis. “Marvell’s ‘The Definition of Love.'” Review of English Studies, new series 6.22 (April 1955): 141-6. JSTOR. Web. 24 November 2009.
  • Davison, Dennis. The Poetry of Andrew Marvell. London: Arnold, 1964. Print.
  • Donne, John. The Complete English Poems. Ed. A. J. Smith. Harmondsworth: Penguin, 1971.
  • Klawitter, George. “Marvell’s Use of John Donne: ‘The Definition of Love.'” APPOSITIONS: Studies in Renaissance / Early Modern Literature & Culture 3 (May 2010). Web. 6 July 2010.
  • Legouis, Pierre. André Marvell: Poète Puritain Patriote 1621-1678. 1928; repr. New York: Russell & Russell, 1965. Print.
  • Legouis, Pierre. Andrew Marvell, Poet. Puritan. Patriot. Oxford: Clarendon, 1965. Print.
  • Leishman, J. B. The Art of Marvell’s Poetry. London: Hutchinson, 1966. Print.
  • Margoliouth, H. M., ed. The Poems & Letters of Andrew Marvell, 2 vols. Oxford: Clarendon, 1927. Print.
  • Margoliouth, H. M., ed. The Poems and Letters of Andrew Marvell. Rev. Pierre Legouis with the collaboration of E. E. Duncan-Jones. 2 vols. Oxford: Clarendon, 1971. Print.
  • Marvell, Andrew. The Poems of Andrew Marvell. Ed. Nigel Smith. London: Longman, 2003. Print.
  • Museum of the History of Science, University of Oxford. Astrolabe Collection. Web. 8 July 2010.
  • National Maritime Museum, Greenwich. The Planispheric Astrolabe. Greenwich: National Maritime Museum, 1976.
  • Nordenskiöld, Adolf E. Facsimile-Atlas to the Early History of Cartography. 1889; repr. New York: Dover, 1973.
  • Oxford English Dictionary Online. Web. 24 November 2009.
  • Schmitter, Dean Morgan and Pierre Legouis. “The Cartography of ‘The Definition of Love.'” Review of English Studies, new series 12.45 (February 1961): 49-54. JSTOR. Web. 24 November 2009.
  • Snyder, John P. Flattening the Earth: Two Thousand Years of Map Projections. Chicago: Chicago UP, 1993. Print.
  • Webster, Roderick and Marjorie. Western Astrolabes. Chicago: Adler Planetarium, 1998.