Origins of Thieves Using Scrolls - Commentary

In the last two postings, I presented archetypal scenes of thieves miscasting spells by (a) Fritz Lieber's Gray Mouser, and (b) Jack Vance's Cugel the Clever. Now for some analysis:


The scenes of Cugel and the Mouser are obviously similar in numerous ways. Both feature roguish, self-aggrandizing characters with minimal training in magic, who with difficulty manage to cast powerful spells, and produce highly unexpected effects. The scenes were both written and published within a span of two years (1964 and 1966). Both character/stories are included by Gygax in the AD&D DMG's "Appendix N: Inspirational and Educational Reading".

Another way in which they are similar is that they aren't using scrolls in the sense of D&D's single-use magic items. Cugel is using a "workbook", and the writing remains afterward for him to double-check and confirm exactly where his mistake was. Mouser is working from a single sheet of parchment, but again, the writing remains afterward (his companion suggests he use it a second time, and his sole counter-argument is how it failed the first time). This isn't too surprising, since to my knowledge, the D&D Magic Scroll is purely an in-game invention to provide a single-shot magical effect. I can't think of any source prior to D&D that features an item exactly like that, with disappearing writing once the magical charge is released (can you?).

So these are really just "regular" spells in each case, in that the rogues are forced to memorize them prior to casting -- Cugel has "established the activating sounds carefully within his mind" (in the canonical Vancian sense), and the Mouser must keep "eyes tightly shut while the last sentences of the rune were being recited and the great forbidden words spoken; even the tiniest blink would nullify the Great Spell."

A difference: In Cugel's case, the story is explicit in that the surprise magical effects (imprisonment reversed to freedom; transporting enemies reversed to transport of self) are definitely due to misspeaking syllables on the caster's part ("Cugel hastily consulted the workbook and saw that in error he had transposed a pair of pervulsions, thereby reversing the quality of the spell."). But in Vance's Gray Mouser story, the truth is left ambiguous. Possible causes suggested in the text are that (a) the Mouser may be "doing the spell backwards", (b) the sorcerers of Gwaay were lying about being of the "First Rank", or (c) "Sheelba had underestimated its power". Option (a), a miscast by the Mouser may in fact be the most reasonable (destruction of sorcerers other than the First Rank, reversed to destruction of sorcerers only of the First Rank) -- even though the text goes so far as to directly state that "he finished without flaw" [p. 729], this may be understood as coming from an untrustworthy narrator, directly on the heels of "He almost stuttered midway through the word 'slewerisophnak'". It seems unlikely that Gwaay's less-numerous sorcerers are weaker, lying, and that simultaneously the Mouser is himself an undiscovered sorcerer of the First Rank (as he tries to rationalize at one point) -- although it seems equally unlikely that Sheelba would give him a spell expected to cause his own destruction when normally cast. In summary: the situation with the Mouser is left murky.

Personally, I would prefer to simulate these scenarios via multi-classing (as opposed to a unique thief ability with scrolls), saying that Cugel and the Mouser had some low levels in magic-use/wizardry, and that low-level wizards could try to use high-level spellbooks with some significant chance of failure. That seems a truer representation of what's really happening in these cases; although admittedly it's not a mechanic that otherwise exists in classic D&D (excepting the kernel of a system in Chainmail itself for delayed or non-effective spells).

Note that in the history of D&D, scrolls came first (in the OD&D little brown books, Vol-2), and thieves appeared second (in OD&D Supplement I, Greyhawk). So in some sense the thief ability could be grafted onto the pre-existing one-shot magic item, without giving them possibly unbalancing access to any wizard's full spellbook (and the same goes for other low-level wizards). If the timing of the development had been reversed, I doubt we'd see this mechanic exactly as we do.

Finally, one thing that always bothered me about scrolls in AD&D (and truthfully, drove this particular investigation) is the question: If scrolls are purely magical words on paper, and lack any somatic (gestural) or material components, then how can wizards transcribe scroll-spells into fully memorizable book-spells? (That is, presumably they would lack the usual hand-gestures, et. al.) But one solution I can see is to roll back to the OD&D understanding (as is often the case). OD&D has no requirements for somatic or material components. Likewise, in neither Vance nor Leiber are arcane gestures emphasized -- in each case, it is the words of power alone that are highlighted, the supernatural syllables that must be carefully formed and do the work (even with the caster's eyes closed; possibly with an exception of simply naming and pointing at the target). So if we synthesize these trio of observations and say that magic spells are spoken words only, then whatever is read from a magic scroll is the same as the full spell formula itself (again, possibly substituting a target's name, etc.), and it makes sense to transfer from one to the other, in both directions. We would have to discard the Gygaxian elaborations in AD&D as being an unnecessary and possibly contradictory development, but then it wouldn't be the only case of that.

Poll question -- Does it make sense to transcribe scroll-spells to book-spells? (Results here.)


  1. I had not previously thought about the issue of somatic and material components as they relate to scrolls, but you raise a very good point. Since both scrolls and spells are purely verbal means of spellcasting in OD&D, it seems likely that Gygax simply carried over the original mechanics for scrolls without thinking about their contradictions with the new component rules for memorized spells. I have never been too fond of the component system, especially when it comes to material components; it introduces too much fiddly bookkeeping without improving play. (Admittedly, expensive or rare material components can act as a brake on the casting of some powerful spells, but this doesn't seem an adequate recompense for having to keep track of various little bits of stuff which the magic-user must have secreted somewhere about his person.) Your observation provides a good reason to remove a feature which I never liked and only partially and inconsistently used.

  2. When I first read the Lieber story lo these many years ago, I took it that Gwaay's sorcerors weren't as high ranked as they were cracked up to be, but a transposing on the Mouser's part seems to make more sense on a closer reading.

    I think one of the things we can take from this is latter day, very crunchy D&D editions don't give the GM as much lattitude to mess around with spell effects like Vance and Lieber's miscastings.

  3. Mouse *did* have magical training - in his origin story, he's studying under a sorcerer with indifferent success.

    As for the scroll/spell question, I like the idea that spellcasting requires motion, and it can be rationalized as part of why scrolls burn out - the writer has to translate the gestures into mystic runes, and it is these runes that 'burn out' when the scroll is used. (That also helps explain why thieves can use but not write scrolls, and why spellcasters can fail to learn from a captured scroll - they have to translate back from 'runes' to 'gestures', and it's not easy!

  4. (That should be 'Mouser', not 'Mouse' - though he goes by Mouse in the story I'm referring to if I recall correctly.)

  5. I guess one logical extension would be that spells exist only in one form at a time. On a scroll, in a spell book, or in the casters mind ready to be cast.

    Perhaps it is what makes spell books so valuable. The caster memorisers the spell, and it disappears from the spell book. Only to reappear the next sunrise, or after the caster has cast the spell, if that is preferred.

    Whilst scrolls once spoken, disappear forever.

  6. I don't like the idea of ditching the requirement that casters must be able to both speak and gesture to cast spells. However, I am inclined to ignore all the complicated stuff in 1e about "somatic" components and just assume that the caster needs to do some fairly simple hand waving and pointing.

    As far as scrolls go, my initial thought is that spell scrolls include instructions for what gestures to use as well as the magic words that must be uttered.

  7. Good comments here -- Personally I do most fully agree with John and BigFella near the top of the thread.