THREE == SPECIAL CONSTRUCTIONS

*3.1 == Word-grafting*
*3.2 == i.zaafat constructions*
*3.3 == o constructions*
*3.4 == al constructions*

So far every kind of syllable pattern we've considered has existed within the boundaries of a single word. But several special constructions can generate syllables that ignore word boundaries.

3.1 == Word-grafting

Word-grafting is our term for an operation which the poet may choose to perform on any suitable pair of adjacent words in a line of poetry. The words are suitable if and only if the first word ends with a consonant, and the second begins with a [alif] or aa [alif madd].

Word-grafting consists of pronouncing the two words as though they were run together into one single long word, and scanning them accordingly. If you have trouble performing word-grafting by merely altering the pronunciation of the two words, you can duplicate the same process orthographically by writing the second word without its [madd]--if a [madd] is present--or without its [alif] entirely, if a [madd] is not present. After this initial shortening, the rest of the second word is written as though it were a continuation of the first word. The resulting long word is scanned normally. Here are some examples, covering a range of metrical possibilities:

aa;xir is [aa-;xir is], normally (= = =), can be treated as though it were aa;xiris [aa-;xi-ris], and scanned (= - =)

aap agar [aa-p a-gar], normally (= - - =), can be treated as though it were aapagar [aa-pa-gar], and scanned (= - =)

aa;xir agar [aa-;xir a-gar], normally (= = - =), can be treated as though it were aa;xiragar [aa-;xi-ra-gar], and scanned (= - - =)

aap aa;xir [aa-p aa-;xir], normally (= - = =), can be treated as though it were aapaa;xir [aa-paa-xir], and scanned (= = =)

The effect of word-grafting is always to cram more words into a given amount of metrical space, either by reducing the number of syllables they are divided into, as in the second and fourth examples, or by replacing a long syllable with a short one, as in the first and third examples.

Word-grafting is one of the poet's subtlest and most versatile tools. Though it alters the pronunciation of the words involved and transforms their scansion, it never changes their orthography on the printed page. The presence or absence of word-grafting can be determined only by careful analysis of the metrical environment in which the relevant word-pair occurs. But it certainly occurs less than half the time, so the first reading of a line in an unknown meter cannot assume it.

Two words in succession, or even three or four, may be grafted, and the metrical change can be quite dramatic: Ghalib's kaafir in a.snaam [kaa-fir in a.s-naa-m], normally (= = = = = -), can be treated as kaafirina.snaam [kaa-fi-ri-na.s-naa-m] and thus can be scanned (= - - = = -).

SPECIAL CASES: Word-final ii , e , o may sometimes be pronounced and scanned as consonants to permit word-grafting; but this is very rare. Even rarer is the treatment of (( [((ain] as though it were a [alif] in order to permit word-grafting, as in Mir's ;xaak ((anbar [;xaa-k ((an-bar] which instead of its normal (= - = =) is treated in one poem as [;xaa-k((an-bar], (= = =). Such liberties are no longer taken.

3.2 == i.zaafat constructions

An i.zaafat is a grammatical construction borrowed from Persian that when placed between two nouns, makes the second modify the first; when placed between a noun and an adjective, it affirms their mutual relationship. It is commonly used both with Persian nouns and names and with Arabic ones as well. Its metrical behavior varies according to the last letter of the word on which it is placed.

AN [i.zaafat] ON A CONSONANT: When an [i.zaafat] is applied to a word ending in a consonant, it joins with the last letter of the word to form a flexible syllable. For example:

lab [lab] with an [i.zaafat] is scanned as [la-be], (- x)

mulk [mul-k] with an [i.zaafat] is scanned as [mul-ke] (= x)

diivaan [dii-vaa-n] with an [i.zaafat] is scanned as [dii-vaa-ne], (= = x).

But it causes no other changes in the word's scansion, even in three-letter three-consonant words where such changes sometimes occur. Consider the case of na:zar [na-:zar], "sight," which is scanned (- =). Its direct plural form is na:zreN [na:z-re;N], (= x); its oblique plural form is na:zro;N [na:z-ro;N], (= x); but with an [i.zaafat] it is scanned [na-:za-re], (- - =). In this latter case the final syllable cannot be flexible, because three shorts can never occur in a row; see Chapter 7.

ARABIC MONOSYLLABIC WORDS of the form "consonant + consonant" sometimes have a special form that violates this rule. The application of an [i.zaafat] to such words produces either an optional [tashdiid], or with some words even a compulsory [tashdiid], on the second of the two consonants. For example, fan [fan] with an [i.zaafat] becomes either [fa-ne], (- x), or [fan-ne], (= x), with a [tashdiid] over the n ; and ;xa:t [;xa:t] with an [i.zaafat] becomes either [;xa-:te], (- x), or [;xa:t-:te], (= x). (Sometimes, though not so often, the same thing is done before a conjunctive vaa))o [-o-].)

There's no simple way to decide, on seeing a word of this kind, whether it must, or simply might, have the [tashdiid]. The most convenient way to allow for this effect is therefore initially to scan all Arabic two-consonant words followed by an [i.zaafat] as "flexible"-flexible; though technically speaking only the second syllable is a flexible one.

Apart from this case of two-consonant Arabic words, the rule for [i.zaafat] on words ending in consonants is never broken.

AN [i.zaafat] ON AN [alif]: When an i.zaafat is applied to a word ending in the letter a [alif], in modern orthography the letter e (that is, a ba;Rii ye ) is added as a symbol of the [i.zaafat], and the [i.zaafat] constitutes one flexible syllable. The [i.zaafat] may be indicated by the letter e alone, or by the letter e with a )) [hamzah] above it, or (incorrectly) by a [hamzah] alone, or, in some older books, by a [zer] alone. All these forms are scanned identically. Moreover, in such cases the word-final syllable ending in a [alif] is never flexible, but is always scanned as LONG. Thus vafaa [vafaa], when followed by an [i.zaafat], becomes [vafaa-e], (- = x).

AN [i.zaafat] ON A o : When an i.zaafat is applied to a word ending in o pronounced as a vowel, usually it is treated the same way as in the case of a [alif]. It thus receives the letter e to represent the [i.zaafat], and the [i.zaafat] forms one flexible syllable. For example, kuu [kuu] followed by an [i.zaafat] becomes [kuu-e], (= x). Most often (though not always) this isolated [i.zaafat] syllable is short.

In some cases, however, the application of [i.zaafat] to word-final vowel- o causes that o to be pronounced and scanned as a consonant- o . This usually happens when the vowel- o has the sound of [au], as indicated by a [zabar]. For example, jau [jau] with an [i.zaafat] becomes [ja-ve], (- x); .zau [.zau] with an [i.zaafat] becomes [.za-ve], (- x).

AN [i.zaafat] ON ii : When an i.zaafat is applied to a word ending in ii pronounced as a vowel (that is, chho;Tii ye ), as a rule the [i.zaafat] causes the word-final vowel- ii to be pronounced and scanned as a consonant , and a normal consonant [i.zaafat] is formed. For example, sho;xii [sho-;xii] with an [i.zaafat] becomes [sho-;xi-ye], (= - x); dushmanii [dush-ma-nii] with an [i.zaafat] becomes [dush-ma-ni-ye], (= - - x).

Sometimes, however, the poet may treat the word-final ii as a full vowel, and give it an entirely separate [i.zaafat]-syllable like those given to a [alif] and vowel- o . In these rather rare cases the word itself terminates in a LONG syllable, and the independent [i.zaafat]-syllable is almost always short. For example, Atish writes saaqii-e azal [saaqii-e azal], making [saa-qii-e] scan (= = -), rather than the usual [saa-qi-ye], (= - x). An example from Mir: M{932,2}.

AN [i.zaafat] ON e : When an i.zaafat is applied to a word ending in {e} pronounced as a vowel, the application of the [i.zaafat] causes the vowel- e to be pronounced and scanned as a consonant, and a normal consonant [i.zaafat] is formed. For example, mai [mai] with an [i.zaafat] becomes [ma-ye], (- x).

3.3 == o constructions

As an independent word, o (that is, the letter vaa))o ) means "and." In Urdu it is normally pronounced as a long vowel [o]. It is borrowed directly from Persian grammar, and occurs only between two Persian--or sometimes Arabic--words or proper names. Its behavior is in many ways similar to that of [i.zaafat]. When [o] follows a word ending in a consonant, it joins with that consonant to make one flexible syllable: diin o dil is scanned as [dii-no dil], (= x =). This is the normal pattern. And just as with [i.zaafat], Arabic two-letter two-consonant words sometimes receive a [tashdiid] on the word-final consonant before the [o]: ;xa:t [;xa:t] followed by o might become either [;xa-:to], (- x), or [;xa:t-:to], (= x).

o FOLLOWING a [alif]: When o follows a word ending in a [alif], it always forms an independent syllable by itself. This independent syllable is usually short, and the word-final syllable before it, ending in a [alif], is always long. Thus vafaa [vafaa] with o usually becomes [va-faa o], (- = -) though at times it might be treated as (- = =).

o FOLLOWING ii : When o follows a word ending in ii (that is, the letter chho;Tii ye ), usually the same thing happens: it forms an independent syllable. This independent syllable is usually short, and the word-final syllable before it, ending in ii , is always long. Thus saadagii o [sadaagii o] normally becomes [saa-da-gii o], (= - = -), though (= - = =) may also occur.

Sometimes, however, it may happen that the o turns the vowel- ii into a consonant, so that a normal consonant- o construction occurs: the ii and the o together form one flexible syllable, and the syllable before it is thus always reduced to a one-letter short one. Mir occasionally does this sort of thing: badnaamii o [badnaamii o] is scanned as [bad-naa-mi-yo], (= = - x), or shaadii o [shaadii o] as [shaa-di-yo], (= - x).

o FOLLOWING e OR o : When o follows a word ending in e or o , word-final vowels are usually turned into consonants, and scanned as such. Having become consonants, they join with o as consonants normally do: mai o [mai o] becomes [ma-yo], (- x); ;xusrau o [;xusrau o] becomes [;xus-ra-vo], (= - x). Note that if the word-final o is already a consonant, it readily behaves as the other consonants do: sarv [sarv] when followed by o becomes [sar-vo], (= x).

o FOLLOWING h OR ;h : When o follows a word ending in h , it usually behaves in the consonant pattern, joining with the h to form a single flexible syllable. But sometimes the h is pronounced and scanned as a vowel; in this case the o forms an independent flexible syllable. The same range of possibilities exists for ;h .

3.4 == al constructions

The Arabic particle al , usually pronounced [ul] in Urdu, appears between two Arabic words, and unifies them into a phrase. Its relationship with the second of the two words is quite simple: metrically speaking, they are entirely separate. After the al comes a complete break; scansion then begins afresh and proceeds normally. (The distinction between shamsii [shamsii] and qamarii [qamarii] words affects only pronunciation, with scansion remaining the same in either case.)

The word before the al , however, unites intimately with it and is scanned together with it. When the word before the al ends in a consonant, the scansion technique is simple: pretend that the word is written with merely an extra l at the end of it instead of the whole al , and pronounce and scan the word normally. Examples:

((aalam ul-;Gaib} [((aalam ul-;Gaib] is scanned [((aa-la-mul ;Gai-b], (= - = = -)

an al-;haq [an al-;haq] is scanned [a-nal ;haq], (- = =)

lisaan ul-((a.sr [lisaan ul-((a.sr] is scanned [li-saa-nul ((a.s-r], (- = = = -)

((a:ziim ul-shaan [((a:ziim ush-shaan] is scanned [((a-:zii-mush shaa-n], (- = = = -)

When the first word is a two-consonant word, it will always have a [tashdiid] on its final consonant: rabb ul-ra;hiim [rabb ur-ra;hiim] is thus scanned [rab-bur ra-;hii-m], (= = - = -).

In the rare cases when al follows a word ending in a vowel, expect trouble! Assume that orthography will be thoroughly misleading and will not correspond to actual pronunciation and scansion. The reality is invariably shorter than the appearance, but it is hard to formulate general rules since Arabic grammar is the determining factor. Notice the following examples:

[bi] + al = [bil]:

baalkul [bi al-kul] becomes [bil-kul], (= =)

baalaa;xir [bi al-aa;xir] becomes [bil-aa-xir], (= = =)

baaliraadah [bi al-iraadah] becomes [bil-i-raa-dah], (= - = x)

fii + al = [fil]:

fii al-;haal [fii al-;haal] becomes [fil-;haa-l], (= = -)

fii al-faur [fii al-faur] becomes [fil-fau-r], (= = -)

fii al-;haqiiqat [fii al-;haqiiqat] becomes [fil-;ha-qii-qat], (= - = =)

consonant + o + al :

;zuu al-fiqaar [;zuu al-fiqaar] becomes [;zul-fi-qaa-r], (= - = -)

buu al-havas [buu al-havas] becomes [bul-ha-vas], (= - =)

The presence of a )) [hamzah] at the end of the first word, however, prevents this kind of shortening of the vowel before al , as in maa)) al-;hayaat [maa)) al-;hayaat], which remains [maa))-ul-;ha-yaa-t], (= = - = -).