This week's reading is Prosodic Faithfulness and Correspondence: Evidence from a Japanese Argot, by Junko Itō, Yoshihisa Kitagawa, and Armin Mester. The "Japanese Argot" in question is called zūjago, which is itself an example of zūjago: the etymology is zūja "zūjago for 'jazz'" + go "language".

At this point zūjago is sort of a blanket term for "entertainers' argot", but originally it was more closely associated with jazz musicians in particular. All kinds of lexical items get thrown into the zūjago bucket, and it's not hard at all to find wordlists online (in Japanese), but Itō, Kitagawa and Mester are particularly interested in a productive reversing pattern (that Kitagawa in particular apparently "has active control" of), which gives you words like this:

• menfu < fumen 譜面 "sheet music"
• jaamane < maneejaa マネージャー "manager"
• yanopi < piano ピアノ "piano"
• rūbi < biiru ビール "beer"
• gaikichi < kichigai 気違い "crazy"
• patsura < rappa 喇叭 "trumpet"
• kuribitsu < bikkuri びっくり "surprise"

In the IKM analysis, this system works as follows:

a. Reversal: Argot words are reversals of their bases (in a sense to be made precise).
b. Template: Argot words either have the form "F+F" ("foot+foot") or the form "F+L" ("foot + light syllable"), nothing else is admitted.
c. Predictability: Within the limits of (a), the prosodic type of an argot word is determined by the prosodic form of its base.
d. Preservation properties: The mapping relation between base and argot form allows for weight adjustments (mora deletion and mora insertion) and segmental spreading, but not for segment deletion or segment insertion.
e. Null output: As a consequence of (b) and (d), words of Japanese beyond a certain size limit do not have a corresponding argot form.

Rule (a) is pretty self-explanatory, although it's interesting to note how the segmentation works. For example, the gemination marker /Q/, written with (and in many cases etymologically equivalent to) a small tsu, gets reanalyzed as an actual tsu when ends up not followed by a consonant any more, e.g. bikkuri (bi-tsu_small-ku-ri) → kuribitsu (ku-ri-bi-tsu).

The meaning of rule (b) is that a zūjago word must begin with a foot (either a "long syllable" or two light syllables [morae, basically]), and end with either another foot or a single mora.

For example, if you start with sake ("alcohol[ic drink]"), you can't just flip it to kesa, because that gives you two light syllables and nothing else. Instead, you lengthen the first syllable to make it heavy: keesa, a well-formed zūjago word. Similarly, if you start with biiru, you can't just swap the existing parts around to rubii, because this gives you a light syllable followed by a foot (the heavy syllable bii). Instead, the ru gets lengthened into a heavy syllable, and the bii contracts to a light one (for other reasons): rūbi.

A word that is originally just one short syllable, though, like hi ("fire", "light [for a cigarette]"), gets put through the wringer and ends up iihi. The authors see this as working as follows: the initial word hi is analyzed as two overlapping parts, /hi/ and /i/. This is then treated the same way as sake: the syllables are swapped giving ihi and then the new first syllable is lengthened to give iihi, the final zūjago form.

The weakness in a paper like this is obviously going to be that there is no way to produce a rigorous, watertight set of rules for generating the lexicon of a demimonde argot. As an example, one page after the authors declare that original words of six morae or more cannot be zūjagofied (because they can't be processed into a form of two feet maximum), they offer bontoro as zūjago for toronbōn ("trombone"), which is... a six-mora word. However, they do recognize this as an issue; bontoro is, in their analysis (which I presume represents the "native speaker" judgment of Kitagawa), an "isolated and lexicalized" example, not producible by the "productive strategies" they identify. You just have to keep in mind that this paper is a description of a particular corner of a certain language game, rather than a Key to all Jam Session Banter.