Anglo-Romani (Long, was Re: Genetic Descent)

[Vidhyanath Rao]

I had asked, regarding Anglo-Romani as 'relexicalized English', how the
lexicon was  preserved? I think that I need to be more detailed as to
what I was asking. I also took some time to try  to get information on
the language of English Gypsies, and thought the results might be of
interest to some. Before getting to that, I will touch on a different
point I had mentioned, that of syntactic influence.

The paper about Konkan Saraswat Brahmans I had mentioned is Nadkarni,
Biligualism and syntactic change in Konkani, Langauge 51(1975)
pp. 672--683. The inherited distinction between relative pronouns are
being supplanted by interrogative pronouns. [In the dialect of Marathi
spoken around Madurai, Tamil Nadu, this process is complete.] The
conclusion the author comes to is that when two languages are in
contact, with all speakers of one (say A for definiteness) are >early<
bilinguals but most speakers of B are monolingual, and speakers of A
continue to speak A, influence will go from B to A and not in the
reverse. I.e., the direction of influence is determined by the patterns
of social contact and not what language is spoken at home.

Secondly, it seems that in cases where language is used for
self-identification, it is easier for syntax to be borrowed than (core)
vocabulary or morphemes. In fact, sprachbund phenomena are much more
about syntactic convergence than others. Yet we do not consider a
sprachbund to form a new family.

Also, think about modern IA langauges. Only pieces of Vedic inflection
surviving are the oblique (from the genetive) and some bits of the be
auxiliary. The peterite is from the -ta form which was a stative
>adjective<, while other forms come from various auxiliaries. This
wholescale change in morphology is never taken as affecting the getetic
affiliations of ModIA. Why should it be different in cases where the
syntax has been affected profoundly by external influence and not
internal development?

So it seems that the situations like Laha vs. Malay are exceptional
in being complete replacement of syntax, rather than in the process
itself. Now 'morphology', interpreted as 'how to form the words that go
into the various slots recignized by the synchronic syntax' is
dependent on syntax, it will be affected by this process, but that
should not be relevant to genetic classification.

Turning now to Anglo-Romani:

H. W. Hatting wrote, in a message dated May 16, 2001

> Let me put in some information I recollect from a seminar on Romani
> 1. English gypsies were speaking a version of Romani with *Romani*
> grammar when they immigrated into England;
> 2. Gradually, they started to use English among themselves [...]
> 3. Nowadays, English Romani (with Romani lexicon and English
> morphology) [...]
> English Romani goes farther than most other secret languages (like
> German "Rotwelsch" or Russsian "blat", in which the basic function
> words (pronouns, conjunctions, auxiliaries etc.) are normally from
> the "base" language which provides the morphology) [...]

I interpreted this, when combined with the use of "relexification", to
mean that auxiliaries, their conjugation, endings etc continued from
Stage 1 to Stage 3 (with only the categories common to English and
Romani being found) while the speakers in the intermediate stage were
completely ignorant of Romani grammar. That I found, and still find,
incredible. My question as to how the lexicon was preserved was about
bridging the first and third stages. My lack of precision unfortunately
misled JoatSimeon at aol.com who addressed only how Stage 3 can continue
indefinitely, something that is not at all surprising.

Some browsing in the library (which I should have done sooner) suggests
that the intermediate stage was more complicated. "The dialect of
English Gypsies" by Smart and Crofton (1875, reprinted by Gale
Research, 1968, Detroit) talks about two "dialects", termed there "Deep
Romanes" and "Shallow Romanes" or the "vulgar dialect". The former
seems to preserve quite a bit of East European Romani morphemes (S&C
refer to "the Turkish" dialect, presumably the Romani of Gypsies in
(the European parts of?) the then Ottoman Empire), though there were
losses of categories not found in English. The latter dialect seems
closer to Blat etc than Hatting suggested. Here is the beginning of a
story "How Petalengo went to Heaven" retold in both dialects: [Parts
enclosed in angle brackets '<>' are English morphemes, italicized in
print. Acute accent mark on vowels is denoted by a slash '/' preceding
the vowel. In each sentence the old dialect version is given first, the
new second. Smart and Croft do not provide a translation, only a
reference to Hone's "Everyday book", 1857ed, vol. 1, p447. Our library
does not have this edition and I couldn't locate this story in there,
because I didn't know that 'Petalengro' means 'Smith'/blacksmith.]

   Mandi      pookerova    toot sar  Petalengro
   I          tell-nonpast you  how  P.
   Mandi<'ll> pooker tooti <how>     <the> Petalengro

   ghi/as  kater mi Doovelesko keri.
   go-past to    God-possesive house
   jal<d>  adr/e mi Doovel'<s> kair.

[Notes: Doovel (God) is always prefixed with mi (my). Keri is adverb
from kair.]

   Yek divvus mi Doovel vi/as     adr/e bitto       gav.
   One day       God    come-past to    small-masc  town.
   Yek divvus mi Doovel wel<d>    adr/e <a> bitti   gav

   Kek nan/ei    kitchema sas          ad/oi.
   emphatic neg  inn      be-past III  there
   <and> latch<d>  kekeno(?) kitchema od/oi
         find-past NEG (?)   inn      there

   Yov       ghi/as adr/e Petalengro<s>       kair.
   He        went   to    P-poss              house
   <so> <he> jal<d> adr/e <the> Petalengro<s> kair

   Yov   sootada/s       odoi  sor doova raati.
   He    sleep-past III  there all that  night
   <and> sooter<d>       od/oi sor doova raati.

This is striking. The "old" (archaizing?) dialect uses the inherited
inflections, but still slips into 'Petalengro<s>' (because -engro is
already a poss. suffix? (turned into an adjective formant)). The "new"
dialect uses English particles, verb endings (another story, labeled as
being in "the new dialect" has "sas dik<ing>" for 'was looking') etc,
but on occasion retains pronouns (I person always, II seemingly more
often than III) and occasionally auxiliaries. Even Wester Boswell, the
chief informant, considered both by other gypsies of the time and the
authors as having good control of Romani, would slip into the new forms
from time to time, and then correct himself. On the other hand, he
seemed to have preserved curious archaisms: Prohibitions were with
'maw', given as 'maa' in the vocabulary, [e.g., "thou shalt not kill"
is translated as "maw toot maur"], negations with forms of 'kek'
without or without 'na', occasionally the latter by itself. Others seem
to vary between this and the other extreme of the "new dialect": The
comments in Smith and Crofton suggest that the extent of the use of
inherited pronouns and the be-auxillary varied quite a bit. [See "sas
diking" above. A story in a slightly later source (Leland) quoted by G.
Pierce starts "Mandy su:ttod I was pirraben lang o tute" [the ending d
on su:ttod, the words I and was, and probably lang (=along), o (=of)
and en (=ing') are English ], with 'was' instead of 'sas'. Another
story in S&C goes "<A> chooro dinilo jookel sas peer<in'> ... Dikt/as
... Yov piriv<ed> ... adr/e /o paani. ..." Note the mixing of dikt/as,
an inherited peterite, with the English -ed of pirived, and the
articles '/o' (from Greek) and 'the'.

Compounds can also lead to 'mixed' words, sometimes in strange ways: In
the translation of Lord's Prayer, 'forgive' is "<for>del"! [del = give,
cf Mod IA 'de-']

An interesting point here is that when asked for the Romani word for
something, when it had been lost, people would try to come up with some
periphrase: Thus "how do you ask for a spade?" elicited "Lel <the> kova
<to> chin <a> hev adr/e o poov", literally 'Get the thing for cutting a
hole in the ground".  Sometimes, different persons would give different
periphrase. For 'frog', the responses where "tikeni koli <as> jal<s>
adr/e <the> paani, <and> lel<s> <the> drab avr/i" [little thing that
goes into the water and takes the poison out] and "O stor-herengro
bengesko koli ta jal<s> adr/e o paani so piova" [the four legged
diabolic thing that swims in the water which I drink] (this by Wester
Boswell). It seems that the inherited word for frog 'jamba'/'jomba' had
come to mean toad, though it was considered to be a form of 'jumper'.

Thus far the evidence seems to point to language death. The "new
dialect" is what those without adequate control of the ancestral
language spoke, not a "relexification". This is underlined by the fact
that extent of the use of English function words varied from person to
person, with older persons having more control.

The only references about 20th c. Anglo-Romani that I found in our
library are G. Price(ed) "Languages in Britain & Ireland" and Kenrick,
Romani English, in International J Sociology Lang, 19(1979) pp.
111--120. The former seems to be based on the latter for the most part.
I could not find texts in contemporary Anglo-Romani. So I don't know if
it should be considered as an arrested form of the "new dialect", or if
attempts had been made to remove the English function words and
endings. [Apparently, some literate Gypsies have consulted books to
increase their vocabulary.] But what I read suggests that the inherited
inflection is not used.

In any case, contemporary Anglo-Romani looks like a case of arrested
language death or a resurrected language. I am not sure if
"relexification" is a suitable word to use, at least in the sense I
understand it. After all, I have never heard anyone call Sanskrit of
drama dialog as 'Prakrit relexified with Sanskrit' though in syntax the
two are much closer than either is to Sanskrit of Panini/early
upanishads. Even in morphology, there is variation in the amount of
Prakrit influence when we look at the puraanas or Buddhist texts. Long
compounds of medieval Sanskrit are often attempts of reconcile MIA
syntax and rhythm to Panini's morphological rules.

It should be noted that Kenrick who proposed that Anglo-Romani be
considered a register of English (and thus the name Romani English)
still endorsed a view similar to the above about the historical origin
of Anglo-Romani. He just says that considering Anglo-Romani to be a
register of Romani clarifies the current socio-linguistic aspects. This
should not affect how we view its history.

Obviously, dying languages and language shifts in progress are going to
cause fuzzy-wuzzy type problems for genetic classification, just as
sound changes in progress won't conform to the rules used in historical
linguistics. If the process is arrested by "schooling" (of formal or
informal types), as seems to have happened in Anglo-Romani, this can
continue indefinitely. We should not throw out the rules due such
exceptional cases, nor should we attempt to forcibly fit the cases to
the rules without extensive investigation of the history.