y:Üen:v:es:ýXi Aaôf em:eS:g:n:
t:aðVy:að y:Üen:v:es:ýXi Aaôf faõrðn: sXRiz:

Creating learning materials for advanced levels of Hindi-Urdu:
Distinguishing constructions from rules and from idioms

(NOTE:  To download (for free) a version of Netsacpe 4.x that best displays the Xdvng Devanagari font from "SillyDog701", click here.   Then scroll about halfway down to find "Netscape Communicator 4.5 ~ 4.8".   I recommend downloading version 4.8.   This will take about three minutes if using a university computer, longer if using a dial-up connection.)

A. What is an "advanced construction"?

      The term "advanced construction" is intended to encompass those syntactic patterns or structures that are not language-wide rules of grammar but at the same time are not so limited in productivity as to be classed as idioms. Instances of "rules of grammar" are the passive construction (1) or the use of the possessive to indicate the subject of an infinitive (2):

      (1) hm: Vy:a krðø ?    =>     Vy:a eky:a j:aO?
           'What shall we do?'     =>   'What shall be done?'

      (2) Es:s:ð p:hl:ð ek hm: l:aòXðø ...     =>    hm:arð l:aòXn:ð s:ð p:hl:ð ...
           'Before we turn back ...'         =>     'Before our turning back ...'

      Rules of grammar are extremely general, have thousands of possible lexical instantiations, are found in nearly every description of a language, and are almost always covered in the first or second year of formal instruction.
      An "idiom" is a construction limited to only one or two lexical instantiations. Idioms are not usually included in grammars. Often best treated as itself a kind of lexical item, a given idiom may or may not be included in dictionaries. Idioms are numerous enough to preclude any general agreement about which ones are to be included in a language course and which ones are best left out. As an example, consider (3). The idiom in (3) has a single (invariant) lexical instantiation:

      Only some form of the verb (1) haðn:a 'to be' and the noun (1) p:ðX 'stomach, belly' can appear in it.
      (3) (v:h)p:ðX s:ð hò .
           'She is pregnant.'

      What I term "advanced constructions" lie between these two extremes: Less productive (or general) than rules of grammar they are rarely if ever included in grammars. More productive than idioms they are rarely if ever to be found in dictionaries.
      Consider the construction in (4), adapted from a headline in a newspaper (s:anDy: XaEmz:, 9 July 1996):
      (4) daðst: t:að daðst:, b:ap: B:i m:dd dðn:ð n:hiø Aay:a .
           'What to say of his friends, even his father did not come to his aid.'

The construction in (4) can be defined as in (5):

      (5) NPi t:að NPi, NPj B:i X       (where X indicates the remainder of the sentence)
      The formula in (5) indicates that any noun phrase (NPi), occurring twice and separated by t:að, can be used in the first part of the construction, followed in the second part by a different noun phrase (NPj) followed in turn by B:i and finally the predicate. What the formula in (5) does not show is that there must be a hierarchical relationship between the two noun phrases (NPi) and (NPj). In (6) a cop is presumed more likely to be unscrupulous than a lawyer is:

      (6) v:kil: t:að v:kil: p:Øel:es:y:ð B:i S:rif en:kl:ð . (Terry Varma, 18 Nov 1999)
      'What to say of the lawyers, even the police turned out to be gentlemen!'

Another example found by Terry Varma (in a story for children) makes a tongue-in-cheek comparison of men and money on a scale of their ability to disappear in the big city:

      (7) p:aNRð B:Ey:a y:h Ap:n:a g:aúv: n:hiö. m:Øöb:I hò m:Øöb:I j:haú p:òs:ð t:að p:òs:ð Ens:an: B:i K:að j:at:ð hòö. (Terry Varma, 19 April 2000)

          'Brother Pande, this is not our village. This is Bombay. Bombay! Forget about money, even human beings disappear here!'

In (8) a snake outranks a snakeskin on a natural scale of dangerousness; in (9) death outranks Holi in fearsomeness:

      (8) s:aúp: t:að s:aúp:, kñöc:l:i s:ð B:i Rrt:a hò . (Omkar N. Koul, 10 July 1996)
           'What to say of a snake, he's even afraid of a snake's skin!'

      (9) haðl:i t:að haðl:i m:òö m:aòt: s:ñ B:i n:hiö Rrt:a ! (thanks to Terry Varma)
           'Holi? I am not afraid of death itself!'

If the two noun phrases cannot be seen as related and as occupying two successive points on some kind of scale in a way that is natural and relevant to the speaker's intent, the use of the construction in (5) is meaningless. Compare (8) with (10):

      (10) *? kñöc:l:i t:að kñöc:l:i, s:aúp: s:ð B:i Rrt:a hò .
           *? 'What to say of a snakeskin, he's even afraid of a snake!'

      It is semantic considerations such as these that limit the productivity of "advanced constructions" and render them less general than rules of grammar. On the other hand, they are far more productive than idioms: Only the noun p:ðX fits into the idiom in example (3) while the pairs of noun phrases that can be fitted into (5) cannot be listed or enumerated.
      Furthermore, since idioms require the presence of specific content words, dictionaries can be made of them. In contrast, constructions, which are often simply semantic frames defined by function words, cannot be easily arranged or accessed in dictionaries.

B. Advanced constructions and conversational transactions.

      One of my objectives in the Mellon Project is to develop as complete a list as possible of advanced Hindi-Urdu constructions, to find copious examples of them from as diverse a set of sources as possible, to order them in terms of importance, to categorize them in terms of their semantic and discoursal functions, and to identify the limits and conditions on their use. I am doing this by pursuing them in printed as well as "live" sources and by discussing their meanings and productivity with native speakers of Hindi-Urdu and with Indian colleagues in the field of Hindi-Urdu linguistics and language pedagogy.
      One of the striking results that has emerged from this work is the discovery that the constructions collected and characterized so far fall into a limited number of classes defined by their functions in actual or imagined conversations. Some advance the speaker's position in a conversation. The construction in (5) is an example of this. Others of them try to limit the damage to the speaker's position from a point advanced by an interlocutor. These are constructions that concede a point to the interlocutor (who may be only an imagined presence) but which regroup or which reaffirm some crucial point important to the speaker. Consider (11) and (12):

      (11) )D:an:m:n*:i B:l:ð p:Øv:aðü¶:r m:ðö haðö, v:haú kñ l:aðg: n:hiø j:an:t:ð . (n:v:B:art: XaEmz:, 22 May 1997)
           'The PM may very well be in the Northeast; people living there remain unaware of it.'

      (12) y:ad rK:, py:arð; l:aK: Am:rika rhað, t:Ømhari m:øez:l: B:i y:hiø hò
           'Remember, Dear; no matter how long you stay in America, your home is also right here.'

       Examples (11) and (12) are instances of different advanced constructions which are parallel in conversational function (both concede one point then assert or reassert a second) and in form (each has a positive indicator of concession [B:l:ð 'good'; l:aK: '100,000'] and a subjunctive form of the verb). Recognition and analysis of these parallels will permit a more focussed and transparent approach to language pedagogy, one that stresses developing learners' competence in using structures to do things that will help them cope with the challenges of transactional conversation.

Back to Mellon Project Indexpage.

Posted: 19 Aug 1999.
Updated: 31 Aug 1999, 26 Oct 1999, 10 Jan 2000, 25 April 2000, 15 May 2000.