Whatever their origin, the “Para-Romani” languages are specific to service nomads, learned in adolescence (although some may have been spoken natively at some point), and retained as markers of group identity and secret codes. According to Asta Olesen’s Sheikh Mohammadi informants, their children speak Persian until they are six or seven, when they are taught Adurgari, “which is spoken ‘when strangers should not understand what we talk about.’ ” The same seems to be true of the “secret languages” of the Fuga and Waata service nomads of southern Ethiopia.17
When a language foreign to the host society is not available and loan elements are deemed insufficient, various forms of linguistic camouflage are used to ensure unintelligibility: reversal (of whole words or syllables), vowel changes, consonant substitution, prefixation, suffixation, paraphrasing, punning, and so on. The Inadan make themselves incomprehensible by adding the prefix om- and suffix -ak to certain Tamacheq (Tamajec, Tamashek) nouns; the Halabi (the blacksmiths, healers, and entertainers of the Nile valley) transform Arabic words by adding the suffixes -eishi or -elheid; the Romani English (Angloromani) words for “about,” “bull,” and “tobacco smoke” are aboutas, bullas, and fogas; and the Shelta words for the Irish do (“two”) and dorus (“door”) are od and rodus, and for the English “solder” and “supper,” grawder and grupper.18 Shelta was spoken by Irish Travelers (reportedly as a native tongue in some cases) and consists of an Irish Gaelic lexicon, much of it disguised, embedded in an English grammatical framework. Its main function is nontransparency to outsiders, and according to the typically prejudiced (in every sense) account of the collector John Sampson, who met two “tinkers” in a Liverpool tavern in 1890, it served its purpose very well. “These men were not encumbered by any prejudices in favor of personal decency or cleanliness, and the language used by them was, in every sense, corrupt. Etymologically it might be described as a Babylonish, model-lodging-house jargon, compounded of Shelta, ‘flying Cant,’ rhyming slang, and Romani. This they spoke with astonishing fluency, and apparent profit to themselves.”19
Various postexilic Jewish languages have been disparaged in similar ways and spoken by community members with equal fluency and even greater profit (in the sense of meeting a full range of communicative and cognitive needs as well as reinforcing the ethnic boundary). The Jews lost their original home languages relatively early, but nowhere—for as long as they remained specialized service nomads—did they adopt unaltered host languages as a means of internal communication. Wherever they went, they created, or brought with them, their own unique vernaculars, so that there were Jewish versions (sometimes more than one) of Arabic, Persian, Greek, Spanish, Portuguese, French, and Italian, among many others. Or perhaps they were not just “versions,” as some scholars, who prefer “Judezmo” over “Judeo-Spanish” and “Yahudic” over “Judeo-Arabic,” have suggested (echoing the “Angloromani” versus “Romani English” debate). Yiddish, for example, is usually classified as a Germanic language or a dialect of German; either way, it is unique in that it contains an extremely large body of non-Germanic grammatical elements; cannot be traced back to any particular dialect (Solomon Birnbaum called it “a synthesis of diverse dialectal material”); and was spoken exclusively by an occupationally specialized and religiously distinct community wherever its members resided.20 There is no evidence that the early Jewish immigrants to the Rhineland ever shared a dialect with their Christian neighbors; in fact, there is evidence to suggest that the (apparently) Romance languages that they spoke at the time of arrival were themselves uniquely Jewish.21
Some scholars have suggested that Yiddish may be a Romance or Slavic language that experienced a massive lexicon replacement (“relexification”), or that it is a particular type of creole born out of a “pidginized” German followed by “expansion in internal use, accompanied by admixture.”22 The two canonical histories of Yiddish reject the Germanic genesis without attempting to fit the language into any conventional nomenclature (other than “Jewish languages”): Birnbaum calls it a “synthesis” of Semitic, Aramaic, Romance, Germanic, and Slavic “elements,” whereas Max Weinreich describes it as a “fusion language” molded out of four “determinants”—Hebrew, Loez (Judeo-French and Judeo-Italian), German, and Slavic. More recently, Joshua A. Fishman has argued that Yiddish is a “multicomponential” language of the “postexilic Jewish” variety that is commonly seen as deficient by its speakers and other detractors but was never a pidgin because it never passed through a stabilized reduction stage or served as a means of intergroup communication.23 Generally, most creolists mention Yiddish as an exception or not at all; most Yiddish specialists consider it a mixed language without proposing a broader framework to fit it into; a recent advocate of a general “mixed language” category does not consider it mixed enough; and most general linguists assign Yiddish to the Germanic genetic group without discussing its peculiar genesis.24
What seems clear is that when service nomads possessed no vernaculars foreign to their hosts, they created new ones in ways that resembled neither genetic change (transmission from generation to generation) nor pidginization (simplification and role restriction). These languages are—like their speakers—mercurial and Promethean. They do not fit into existing “families,” however defined. Their raison d’ítre is the maintenance of difference, the conscious preservation of the self and thus of strangeness. They are special secret languages in the service of Mercury’s precarious artistry. For example, the argot of German Jewish cattle traders (like that of the rabbis) contained a much higher proportion of Hebrew words than the speech of their kinsmen whose communication needs were less esoteric. With considerable insight as well as irony, they called it Loshen-Koudesh, or “sacred language” / “cow language,” and used it, as a kind of Yiddish in miniature, across large territories. (Beyond the Jewish world, Yiddish was, along with Romani, a major source of European underworld vocabularies.)25 But mostly it was religion, which is to say “culture,” which is to say service nomadism writ large, that made Mercurian languages special. As Max Weinreich put it, “ ‘Ours differs from theirs’ reaches much further than mere disgust words or distinction words.” Or rather, it was not just the filthy and the sublime that uncleansed “Gentile” words could not be allowed to express; it was charity, family, childbirth, death, and indeed most of life. One Sabbath benediction begins with “He who distinguishes between the sacred and the profane” and ends with “He who distinguishes between the sacred and the sacred.” Within the Jewish—and Gypsy—world, “all nooks of life are sacred, some more, some less,” and so secret words multiplied and metamorphosed, until the language itself became secret, like the people it served and celebrated.26