A mirror or stone does not erase the people around it. Medieval and early modern records divide communication among an object, a person expected to see, a person who asks or commands, a reported figure, and a scribe. Clm 849 and John Dee’s diary organize those roles differently, but both show why a later summary should not turn “seen in a crystal” into direct speech heard equally by everyone present.
- Medieval witness
- Clm 849
- Recorded action
- Mortlake, 10 March 1582
- Core distinction
- Sight is mediated testimony
Clm 849 groups several technologies of mediated sight
Kieckhefer’s inventory of the Munich manuscript identifies a cluster of divinatory experiments: three versions of the mirror of Floron at folios 37r–40v, operations for discovering a thief or murderer by gazing into a vessel, mirrors of Lilith, a crystal for an uncertain matter, another crystal operation concerning theft, and a fingernail-gazing procedure. Their adjacency makes mediumship a repeated textual problem rather than one isolated recipe. Different surfaces and named figures are fitted to different questions while the compiler preserves them as alternatives.
Several of these operations assign sight to a virgin child. The master prepares or addresses the setting; the child looks and reports. Christian markers such as baptism and virginity are invoked as credentials within the text, which exposes both its clerical language and its unequal human arrangement. Modern scholarship has rightly treated child-medium rites as a history of power over children as well as a history of magic. The manuscript records a prescribed role; it cannot establish that a vision occurred or that every owner performed the experiment.
Dee’s diary separates prayer, sight, hearing, and writing
Whitby’s transcript dates the action to 10 March 1582 at Mortlake. Edward Talbot, later known as Kelley, kneels at Dee’s desk with a framed stone and prays. Dee prays separately in his oratory. Within about a quarter hour Talbot reports seeing a figure in the stone; Dee comes to the stone and asks the figure’s name. The record assigns the reply plainly to Talbot’s hearing even though Dee writes the dialogue. Four roles are therefore present in adjacent lines: observer, questioner, attributed speaker, and recorder.
A later marginal note says the figure was not the true Uriel and points forward to 5 May 1583. The revision matters because it shows that identification did not become stable merely by entering the diary. Event time and annotation time coexist on the page. The instrument bounded the reported perception, Talbot supplied testimony, Dee shaped it into questions and prose, and later interpretation reopened the identity. The action is unusually valuable because its uncertainty is preserved rather than edited away.
Worked comparison: Floron’s child seer and Dee’s divided action
The mirror of Floron sequence begins from a defined divinatory aim: knowledge of past, present, and future. Its variants allocate the optical task to a child medium and attach the named spirit to the mirror. The master’s authority surrounds the child’s looking. This produces a vertical chain: book directs master, master directs child and spirit, child reports what appears. Even without repeating the operation, the sequence reveals why the mirror is more than a prop. It assigns a place where information is expected to become visible and a particular person through whom that information must pass.
Dee’s action distributes the work more visibly in the record. Talbot positions the framed stone and reports sight; Dee withdraws to pray, returns, questions, and writes; the answer is marked as heard by Talbot. The later note then judges the named figure. The stone never becomes an independent narrator. Compared with Floron’s prescriptive experiment, Dee’s diary reads as an event log with shifting voices and retrospective correction. One tells a reader how roles should be arranged; the other shows how those roles were represented in a dated action.
The bounded conclusion is methodological. A statement such as “the spirit said” compresses at least three transfers: a figure is reportedly perceived, a medium reports words, and a recorder turns the report into text. Clm 849 and Sloane MS 3188 belong to different traditions and address different beings, yet both depend on mediated testimony. Their comparison helps read demon-communication sources without either asserting contact or dismissing the records’ internal organization.
Objects and witnesses should not be collapsed
Clm 849 survives as one paper manuscript with missing opening leaves. Its modern title Munich Manual is editorial shorthand for a miscellany, and Kieckhefer’s numbered experiments are a scholarly map rather than medieval chapter labels printed at the top of every operation. Sloane MS 3188 is Dee’s conference record, with missing leaves and later copying represented elsewhere. A modern transcription makes both easier to navigate while introducing its own headings, punctuation, and decisions.
Surviving crystals and mirrors associated with historical practitioners form another evidence class. Their material can be measured and their ownership investigated, but an object cannot reveal which words were once reported through it. Conversely, a diary’s phrase “stone in the frame” cannot identify every later museum object called Dee’s. Preserving those boundaries is part of reading the communication scene: text describes roles, an artifact preserves matter, and a later edition supplies access. None by itself supplies the whole event.