29 Reducing an element to zero: lopa

Sutra 1.1.60 introduces a concept that has, if I remember right, been made much of in linguistics theory: the ‘zero’ affix or element, which still remains in a hidden form to condition the preceding base. This concept allows many transformations to be carried through in a virtual sense even without an actual element being present, somewhat like scaffolding and centering allows complicated constructions to be erected.   

1.1.60 adarśanam lopah

a-darśanam (1/1) lopah (1/1)

This is a definition of the concept lopah, the paraphrase, vŗtti, being:

a-darśanam (‘non-visiblity’) aśravaņam (‘non-audiblity’) anuccāraņam (‘non-vocalizing’) anupalabdhih (‘non-attainablity’) abhāvo (‘non-existence’) varņavināśa (‘letter or element destruction’) iti (thus, and so on) anarthāntaram (‘converting to non-entity, neutralizing’) etaih śabdaih (‘by such words’) yo’artho (‘this meaning’) (lopah)

 “Non-appearance (adarśana) is termed LOPA” (Sharma).

“The substitution of a blank (lopa)signifies disappearance”, also called ‘elision’ (Vasu).

There is a certain amount of (seemingly pedantic!) discussion to the effect that lopa refers to the concept of adarśanam (‘non-visiblity’) etc., and not the actual sounds represented by the letters. In the absence of this clarification in the commentaries, one may be tempted to substitute the actual letters lopa or adarśana wherever a sutra prescribes it. As put by Sharma (Vol.ii, p.61) “…the term LOPA should be  assigned to the meaning of the word adarśana and not to its form, the word adarśana itself”. It is also noted that lopa is the disappearance of some entity that was in existence in some context, and not merely absence.

The grammarians recognise different types of blank or zero elements:

1.1.61 pratyayasya lukślulupah (#60 adarśanam)

Pratyayasya (6/1) lukślulupah (1/1)

The paraphrase, vŗtti is

Pratyaya-adarśanasya (the non-appearance, disappearance of an affix) luk ślu lup iti (like LUK, ŚLU, LUP) etāh samjñā bhavanti (these are the entities, terms)

“Non-appearance of an affix is termed LUK, ŚLU, or LUP” (Sharma, II.62)

“The disappearance of an affix when it is caused by the words LUK, ŚLU, or LUP are designated by those terms respectively” (Vasu, p.56).

These terms LUK, ŚLU, or LUP are obviously technical codes, samjñā, created by the grammarians to trigger certain routines, and not real words in the language (which is why we have shown them in upper case, as we do markers or iT’s). These samjñā will be provided in other rules dealing with elision operations in certain appropriate contexts. There is again discussion (in my opinion somewhat pedantic) of the chicken-or-egg question of anyony-āśrayitva ‘interdependency’  (we saw this in 1.1.45 ig yaņah samprasāraņam, see post #23): there has to be something for the lopa operation to work on, which means there has to be an assignment of the name LUK or ŚLU  or LUP before the elision can take place. “Commentators resolve this difficulty by stating that whenever these terms are used in ordering the operation of zeroing, the assignment of the name follows zeroing. That  is, assignment of a name (samjñā) is treated as ‘yet to be brought about’ (bhāvinī)” (Sharma, II.63). The three types of elision (zeroing) operations are collectively termed LU-mān, and are marked respectively by K, Ś and P in the connected rules. It is noted that the LU- elisions are applied to affixes (pratyaya) and not to bases.

Here are some examples from Vasu (p.56). Rule 2.4.72 attaches the label LUK to the affix (augment, vikaraņa) śap in forming the present singular from the verb ad, ‘to eat’: ad + śapLUK + ti = ad+ti = atti ‘(3^rd person/1) eats’. The LUK indicates that the śap (which would have introduced –a-) is elided. In forming juhoti ‘(3^rd person/1) invokes’ the śap is elided by the label śLU (2.4.75). Why the same code is not used in both situations is not clear at this point, hopefully it will become clearer as we come across those other sutras.

Here is a further complication or extension dreamed up by our grammarians: certain operations are carried out even under the elision condition (this is the feature so admired by twentieth century linguistics). This gives the grammarian the facility to make general rules of transformation applicable in a wider range of situations.

1.1.62 pratyayalope pratyayalakşaņam

Pratyaya-lope (7/1)  pratyaya-lakşaņam (1/1)

Vŗtti, paraphrase:

Pratyaya-lope kŗte ([even] given affix-elision) pratyaya-lakşaņam (affix-? Indication; “the effect by which an affix could be recognised”, Vasu p.57) pratyaya-hetukam kāryam (operations conditioned by affix) bhavati

“An operation conditioned by an affix applies even if the affix has been replaced by zero” (Sharma).

“When elision of an affix has taken place (lopa), the affix still exerts its influence, and the operations dependant upon it, take place as if it were present” (Vasu).

The repetition of the words pratyaya and lopa indicate that this rule obtains only where the whole affix has been elided, and not if only a part of it has been elided. Only such subsequent operations are allowed as are caused by the affix as such, not orthographic changes due to particular letters etc.

As an example is the word agnicit ‘he who heaped the fire’ or somasut ‘he who pressed soma (the energetic drink of the Vedic priests)’, which have had their final case ending sUP (nominative case, singular, see the Vibhakti Page) elided or lopped off (lopa) . A pada is defined as an entity with a case ending sUP-anta or a verb ending tiŊ-anta (1.4.14 suptiŋantam padam). On this basis, the words cited will not be called pada, so that operations permitted on a pada will be precluded. However by the sutra 1.1. 62 the entity still functions like a pada, as if the elided case ending sUP or  affix (pratyaya) were still visible.  

As is usual, the latitude conferred by #62 is restricted by

1.1.63 na lumatāŋgasya

Na (0)  lumatā (3/1) aŋgasya (7/1) (#62 pratyayalakşaņam)

Vŗtti:

lumatā śabdena (‘by a term which contains [the marker] LU’) lupte pratyaye (given the elision of an affix’) yad aŋgam (‘a pre-suffixal base’) tasya (‘its’) pratyaya-lakşaņam kāryam (‘operation in the presence of the affix’) na bhavati (does not take place’).

“An operation for a pre-suffixal base (aŋga) does not apply (although it would otherwise apply by 1.1.62) if the affix in question is replaced by zero referred to by a term which contains LU (LUK, ŚLU or LUP; see 1.1.62)” (Sharma).

Thus not all elisions to zero preserve the original character of the lopped off (lopa) affix. If the lopping has been the end effect of a LU instruction, then the elided affix’s properties will not persist. The definition of an aŋga, or pre-suffixal base, is the root (dhātu) followed by the affix (pratyaya), before attaching the case-ending (sUP).

28 Substitute treated like the original

The next bunch of sutras talks of the equivalence of the substitute with the original (substituendum) for subsequent operations, with certain qualifications or exceptions. That is, we can do unto the substitute  as we would to the original, under certain conditions.  The first of these sutras is

1.1.56  sthānivad ādeśo’ nalvidhau

The word parsing is as follows:

Sthāni-vat (0) ādeśah (1/1) an-al-vidhau (7/1)

The first term means ‘like the Sthāni, i.e. the original’; being an indeclinable, it is marked (0) as to case and number. The second word, ādeśah, means the ‘substitute’; in the general parlance it means ‘an order, direction’, and is in nominative case, singular number (1,1). The third word is made up of the negative prefix an-, then aL, which is the pratyāhāra or code standing for ‘all letters’ (click on the Pratyahara TAB at top of page for some more info on these!); and vidhau, the seventh case (locative singular, 7/1) of vidhi, which means ‘rule’. The word al-vidhi is ‘a rule (based on) a letter’, or say a ‘letter-rule’; an-al-vidhi would be the ‘negation‘ or ‘absence’ of such a rule. The locative case form an-al-vidhau can be taken to mean ‘in the absence of a letter-rule’.

The paraphrase,  vŗtti, is:

Sthāni-vat (like a substituendum, i.e. the original) ādeśah (substitute) bhavati (is), āśrayeşu kāryeşu (? in related operations), an-al- āśrayeşu (? in non-letter related) sthāni-al-āśrayāņi (?letter-related) kāryāņi (operations) varjayitvā (having excluded)

“A substitute is treated like a substituendum, except when an operation relative to an original sound (aL) is to be performed” (Sharma, Vol.II, p.56).

“A substitute (ādeśa) is like the former occupant (sthānī), but not in the case of a rule the occasion for the operation of which is furnished by the letters of the original term” (Vasu, p.42).

We understand by this that any (subsequent) operations can be performed on the substituted term just as if it were on the original, except where these subsequent operations are closely linked to the actual letters or types of letters that were replaced. For example, if a vowel were to be replaced by a non-vowel, operations specifically possible on the vowels would obviously not be carried over to the substitute non-vowel. Let’s look at some of the examples provided; here Vasu is the more helpful, as he discusses each type of substitute individually.

According to the śāstrakāra-s then, there are 6 types of substituenda (plural of substituendum, (sthānī): dhātu ‘root’, aŋga ‘pre-suffixal base’, kŗt ‘primary affix’, taddhita ‘secondary affix’, avyaya ‘indeclinable’, sUP ‘nominal ending’ or ‘case affix’, tiŊ ‘verbal ending’ or ‘conjugational affix’, and pada ‘fully inflected word’ (Sharma, p/57; Vasu, p.43). In each case, the substitute orderd by any rule, would get the same treatment as the original it replaces, whatever the type was: substitute of a root (dhātu) is treated as a root, substitute of a case-ending (sUP) is treated as a case-ending, and so on. Let’s take just one example, from the many discussed by Vasu.  

From the table of noun case-endings or sUP (click on ‘Vibhakti’ TAB at top of page to get a tabular statement!), we know that Ŋe (Ŋ is the nasal at the back of the soft palate, ng) is the code for the dative (fourth, caturthī) case, denoting ‘to’ the noun. Obviously the letter Ŋ is just a marker or code provided by the grammarian (which is why it is shown in capitals by convention), and the actual case affix is –e. The marker Ŋ also shows that it is of the type ŊiT (see sutra 1.1.53), which is a grammarian’s contrivance, probably so that it can attract certain other rules with their own specific consequences. Now, as per the example cited by Vasu (p.44), by 7.1.13  ‘ya is the substitute of the sUP-affix Ŋe after an inflexive base ending in short a’. Then the ya is treated just like the original Ŋe, and by another sutra 7.3.102, there is “lengthening of the vowel”, e.g. vŗkşāya, dative of vŗkşa ‘to the tree’.

Now we need an example for the clause na-al-vidhau, ‘not in the presence of a letter-rule’, i.e. if the transformation refers to the particular sound (aL) in the original. Vasu (p.45) gives very briefly the following examples of substitution: of –v by –au in the word div by rule 7.1.84 to give inflected case-form dyauh, -n by -ā in pathin by 7.1.85 to give panthā, -d by –a in tvad by 7.2.102 to give the sah. The point is that of these affixes had been treated just like the originals. Rule 6.1.68 would have applied, “and the case-affix su [sU, the nominative singular] would have been elided”.  

Sharma (Vol.II) gives detailed examples in the Appendix (something which I have only recently realised!), and the an-aL-vidhau rule is illustrated on p.397 onwards in each of four distinct interpretations of the term. Vasu’s examples come under the second variation, “alah vidhih ‘operation obtaining after an item occurring after aL’ ” (Sharma, p.398). As rule 7.1.84 div aut (#1.1.52 alo ‘ntyasya) would require replacement by au of the terminal letter –v, Shrma says that the final s would have to be deleted in the inflected form div + s (U) = diau + s (I am not quite clear why!). Instead, we assume that an-al-vidhau comes into operation, the au is not treated as a v, so that diau+s will yield dyau+s, and by conversion of s into visarga h, dyauh. Similarly or the other examples above.

There are two more sutras in this vein:

1.1.57 acah (6/1) parasmin (7/1) pūrvavidhau (7/1) (#56 sthānivad ādeśah), whose paraphrase is:

Ac ādeśah (aj ādeśah) (substitute for a vowel) paranimittakah (? Conditioned by a right context) pūrvavidhau kartavye (with respect to an operation on a preceding element) sthānivad bhavati (is treated like its substituendum)

Further exceptions to 1.1.57 are given in 1.1.58 and 1.1.59, which I will not go into here at this first reading!

27 Substitution: some ‘prior exceptions’

We had skipped over a couple of sutras in taking up 1.1.49 on substitutions introduced by the sixth case (post #24 to #26). This is as opportune a juncture as any to dispose of these, which are in the nature of ‘prior exception’ to the locational intructions for substitution already covered in the afore-cited posts. We start with

1.1.46 ādyantau ţakitau

Ādi antau, ādyantau (1/2) ţakitau (1/2)

Both are in nominative case, dual number. The first word means ‘first and last’, ‘beginning and end’;  the second term refers to two types of entities (affixes, for instance) marked by the letters Ţ and K respectively (these markers are supplied by the grammarian when giving labels to the respective affixes; they are not something natural or intrinsic to the language!). The sutra lays down that affixes marked with Ţ will attach to the start of the substituendum or sthānin (1.1.49 şaşţhī sthāneyogā, see post #24), and those marked by K will attach to the end. This sutra overrides any subsequent ones, such as 1.1.52 alo’ ntyasya, which fix the location of the substitute at the end of the original. This sutra 1.1.46 (as also #47, #48) is termed a prior exception to 1.1.49, which is why it is given before the latter.

The paraphrase,  vŗtti, is:

 Ādis Ţid bhavati (the initial is marked by  Ţ, i.e. is a Ţit), antah Kid bhavati (the final is marked by  K, i.e. is a Kit), şaşţhī nirdişţasya (of that specified in the genitive, sixth, case). We recall that iT is the technical term for generic marker terms.

“Elements marked with Ţ or K become the initial and final segments respectively of items in the genitive  to which they are introduced” (Sharma, Vol.II, p.45).

“ Of whatsoever the augments enunciated are distinguished by an indicatory Ţ or K, they precede or follow it accordingly” (Vasu, p.35) (a somewhat clumsy expression?).

The point to note is that this does not prescribe substitution of the original like 1.1.49 onwards, but an addition to it, either at the start or the end, respectively. Such elements are termed vikaraņa, augments, rather than ādeśa, substitutes (substituents) (Sharma, p.46). The elements are prescribed in other rules, naturally.

Examples given in Vasu (p.35) are easier to follow, since he gives the relevant sutra number, although the same examples are also cited in Sharma (p.46). One is sutra 7.2.35, which says that an ‘ārdhadhātuka’ affix starting with any consonant other than y (valādi, which leaves out h and y in the pratyahara h^aL of consonants) will take an  iŢ augment. The Ţ marker shows that –i- should be placed at the beginning of the affix.  One such ‘ārdhadhātuka’ affix is the future tense suffix syati, which has to take the –i- augment at the beginning, bhū+iŢ+syati= bhavişyati (Vasu, p.35; obviously other rules determine change of –ū- to -av-, -s- to –ş- etc. Another example cited by Sharma (p.46) is kaņ+tās, where tās is an ‘ārdhadhātuka’ suffix and will take the iŢ augment at the start, kaņ+i+tās = kaņitā.

 An example of an iK augment at the end of an affix (Vasu, p.35) is following sutra 7.3.40, whereby the root bhī (fear) takes the şUK suffix to make the causative (to cause to fear); being marked by K, ş is added at the end of bhī, rather than replacing bhī, giving bhīş-, and the form bhīşayate by further rules.

1.1.47 midaco’ntyātparah

mit (1/1) acah (6/1) antyāt (5/1) parah (1/1)

This is also a prior exception to şaşţhī sthāneyogā, which would have required substitution in its entirety of the word in şaşţhī (sixth case, genitive). This rule states that if the substituting (actually the augmenting) element is marked with M (mIT), it comes in after (parah) the end (antyāt) of the vowels (acah). 

The paraphrase,  vŗtti, is:

Acām sannivişţānām (of the marked vowels) antyād (antyāt) acah paro (parah) (after the end, last vowel) mid (miT) (the augment that has an indicatory M) bhavati (is, comes)

“Given an operational rule which introduces unit Y to unit X, Y should be introduced after the last vowel of X if Y is marked with M” (Sharma, Vol.II, p.46).

“The augment that has an indicatory M comes after the last among the vowels, and becomes the final position of that which it augments” (Vasu, p.35).

Examples will elucidate. Take the augment nUM, where the actual letter to be added would be –n-, the -UM being markers or iT. Thus the derivation of the plural of payas ‘milk’ is, by 7.1.42, payas+nUM+I, where only the n is added to the last of the vowels of payas, which gives paya+n+s+I; the a is lengthened as per other rules (6.4.8), giving payānsi. Sharma gives many other examples.

1.1.48 eca ig hrasvādeśe

ecah (6/1) ik (1/1) hrasva ādeśe (7/1)

 

Vŗtti:

eco (In place of eC vowels: e, o, ai, au) hrasvādeśe (in the case of substitution  by short vowels), kartavya (to be done) ig (ik) (iK  vowels i, u, ŗ, ļ) eva (only) hrasvo (hrasvah) bhavati (are the short vowels) na anyah (not any other).

“If a short vowel is to replace an eC (e, o, ai, au;  and a substitute is not specified), iK (i, u, ŗ, ļ) alone is to be understood as a replacement” (Sharma, Vol.II, p.48).

This talks about the substitution of short vowels for e, o, ai, au, which do not have homologous short versions (Dravidian languages do have short e and short O, for instance). The default substitutes are the iK vowels, i and u, but not really extending to ŗ, ļ as they are far from homologous (the rule 1.1.50 sthāne’ntartamah). It is clear that i is closest to e, and u to o, and these will be the logical replacements for those two long vowels.  In the case of ai and au, ruling out the choice of ŗ and ļ, and also ruling out a (not part of the iK set), the most similar pairs are i for ai, u for au. Thus there are only two logical substitutes i and u, for the four eC sounds (e, o, ai, au). The corresponding pratyāhāra (see relevant Page!) for I and u alone is iŅ, but Sharma guesses that the grammarian did not use this as there is another boundary marker Ņ further down the Shiva-sutras (the sixth, laŅ), which may cause confusion. Of course, that raises the question why the Shiva-sutras use certain boundary markers more than once.

Example:

Vasu gives a simple example following from rule 1.2.47, which declares that “the short vowel is the substitute in the neuter of a crude form provided it end in a vowel” (Vasu, p.36). In the compound ati+rai, the ai must be shortened. The equivalent short being I, we get ati+rai= atiri ‘extravagant’ (Sharma has ‘he who has excelled riches’). Similarly ati+nau= atinu, ‘disembarked or landed’ (Sharma: ‘he who has excelled boating’); upa+go= upagu (‘near a cow’). Avyayībhāva compounds, which are all neuter, will have this replacement. 

26 Further rules on substitution: location and scope

The next few sutras deal with the finer mechanics of the substitution (‘of X, make Y’), eg. the narrowing down of the location where the effect takes place on the substituend (the X):

1.1.52 alo’ ntyasya

Alah (6/1) antyasya (6/1), şaşţhī (from #49)

1.1.53 ŋicca

ŋiT (1/1) ca (0), alah antyasya (from #52), şaşţhī (from #49)

1.1.54 ādeh parasya

ādeh (6/1) parasya (6/1),  alah (from #52), şaşţhī (from #49)

1.1.55 anekāl śit sarvasya

Aneka-al, anekāl (1/1), śiT (1/1), sarvasya (6/1), şaşţhī (from #49)

Let’s start with #52, alah antyasya. The first term is the genitive of al, or the pratyahara aL which denotes the whole range of letters(see the Page on Pratyaharas). The second word means ‘of the last’. By anuvŗtti we supply the understood word şaşţhī  from #49 (and perhaps sthāne). The paraphrase,  vŗtti, with my own translation in brackets, goes as follows:

vŗtti  of 1.1.52

 şaşţhī  nirdişţasya (of the specified sixth case ending, i.e. of the substituend) ya ucyate ādeśah (which is stated the substitute)  sa antyasya alah (that of the last letter) sthāne (in place) veditavyah (? Prescribed, located)

“ A substitute ordered for an item in şaşţhī  ‘genitive’ comes in place of its final aL (‘sound segment’)” (Sharma).

“The substitute takes the place of only the final letter (of that which is denoted by a term exhibited in the genitive or sixth case)” (Vasu). 

That is, the substitute doesn’t displace the whole word. This is extended by the following sutra 1.1.53 to a substitute marked by the label ŋ or Ŋ, the ng sound (which is technically termed a ŊiT); the marker Ŋ will however be dropped in the process of making the substitution):

Vŗtti of 1.1.53, with my attempt at elucidation:

ŋit ca ya ādeśah (also the substitute marked with the letter ŋ, or a ŊiT) sa anekāl (aneka-aL) api (though it may consist of more than one sound or letter aL) alah antyasya bhavati (is of the last letter)  (of the word in genitive, the substituend)

“A substitute marked with Ŋ and ordered for an item in şaşţhī  also comes in place of its final aL ‘sound segment’” (Sharma)

“And the substitute which has an indicatory Ŋ (even though it consists of more than one letter) takes the place of its final letter only of the original expression” (Vasu).

Something to note: these markers are supplied by the grammarian when giving labels to the respective affixes; they are not something natural or intrinsic to the language!

This is an exception (stated in advance!) to #55, which holds that  a multi-letter (aneka-aL, anekāl) substitute will replace the entire word of the original substituend. We will also get back to some sutras that come before this range starting at #49 about substitution, which also amount to an anticipatory exception.

vŗtti  of 1.1.54

parasya kāryam (operation of [on] a following, later [item]) śiśyamāņam (?) ādeh alah (of the starting, first letter, sound segment)  pratyetavyam (?)

“A substitute ordered for a following item comes in place of its initial aL” (Sharma).

“That which is enjoined to come in the room of what follows is to be understood as coming in the room only of the first letter thereof” (Vasu); a decidedly clumsier translation!

vŗtti  of 1.1.55

anekāl (aneka-al, more than one letter) ya ādeśah (the substitute, substituant), śiT ca (and [with] the marker Ś) sarvasya  şaşţhī  nirdişţasya (of all that marked by genitive case) sthāne bhavati (is in place)

“A substitute ordered for an item in şaşţhī  comes in place of the entire item when it (the substitute) consists of more than one aL or is marked with Ś” (Sharma).

“A substitute consisting of more than one letter, and a substitute having an indicatory Ś take the place of the whole of the original expression exhibited in the sixth case” (Vasu).

Thus this bunch of sutras, #52 to #55, are instructions on what is replaced in the X: the ‘default’ seems to be the LAST segment, but if the order is reversed, it may act on the FIRST segment, or if the substitute is multi-syllabic, or it comes with certain markers, it may replace the ENTIRE original.  Obviously these rules can be better understood only with actual examples, Vasu (p.40 onwards) being the more helpful, so here goes.

Examples for 1.1.52 alo’ ntyasya:

Pañcagoņih, ‘purchased for a price of five goņī (sacks)’; similarly, daśagoņih

The rule 1.2.50 id goņyāh (6/1) (see Vasu, p.40) stipulates the replacement by short i in the word that is in sixth case, goņī; this doesn’t mean that the whole word is replaced by short i, only the last syllable (the long ī) is replaced, by 1.1.52 alo’ ntyasya.

Examples for 1.1.53 ŋicca  (alo’ ntyasya):

Mātāpitarau ‘mother and father, two parents’; similarly hotāpotārau ‘two types of officiating priests’

A substitute having more than one syllable should normally replace the entire original substituendum (sthānī) by 1.1.55 anekāl śit sarvasya. However, if the replacement is marked by the indicative letter Ŋ, then it only replaces the initial letter, as in the normal case (#52 alo’ ntyasya). Now compounds formed from the –tŗ roots like pitŗ, mātŗ, hotŗ, potŗ, are governed by the rule 6.3.25 which states  that “In place of words ending in ŗ there is the substitution of ānaŊ in forming dvandvas” (Vasu, p.41), i.e. compounds. This substitute ānaŊ is a ŊiT element, i.e. marked with the Ŋ, so bringing into play rule 1.1.53 ŋit ca, thereby restricting the effect of substitution to the last syllable only. So from words like mātŗ, when used in a compound, only the final ŗ is replaced by ā (I am not quite clear what happens to the –na- in -ānaŊ), thus mātā-pitarau etc.

However, there is an exception to this, in the case of the substitute species tātaŊ in sutra 7.1.35 (Vasu, still on p.41). Even though it has the Ŋ marker, it does replace the whole original -tu and --hi, and not only the final –u and –i. Thus instead of jīvatu ‘may you live’ or jīva ‘may he live’, we get jīvatāt.  Sharma (Vol.II, p.55) however feels that 1.1.53 ŋicca refers to ŋit that do NOT have more than one letter aL, and will therefore replace only the final element by 1.1.52 alo’ ntyasya, while a substitute (ādeśa) that is both marked by ŋ AND has more than one letter, will be governed by the subsequent rule 1.1.55, and will therefore replace the whole substituendum (sthānin).  According to Sharma, the principle which dictates this outcome is rule 1.4.2 vipratiśedhe param kāryam, ‘when in doubt (do) the later operation’, here prefer rule 1.1.55 to 1.1.52. Of course, this does not resolve the case of ānaŊ, unless we argue that it has only ONE letter -ā as the substitute (ādeśa), whereas tātaŊ has more than one, t,  ā, and t. This suggests that we should not be translating aL as ‘syllable’, but stick to ‘letter’ as in Vasu, or ‘sound segment’ as in Sharma.    

Examples for 1.1.54 ādeh parasya (alah from #52, şaşţhī from #49)

dvīpam ‘that which has two waters; island’
antarīpam ‘part of land stretching out into sea’

These are cases where substitution is directed on a word (in sixth case), when it is FOLLOWING some other word: the substitution is made in the word-initial, not in the word-final element (thus an exception to 1.1.52). By 1.1.67 tasmād iti uttarasya (see post #3), this type of ‘context’ is indicated by the ablative (fifth, pañcamī) case for the word or element preceding the substuendum (sthānin, which will be in şaşţhī, genitive or sixth case). One such rule is 6.3.97 dvyantar upasargebhyo apah īt (Sharma, p.55), which is interpreted by Vasu (p.41) as “In the place of ap used AFTER the words dvi and antar and the particles called upasarga, there is ī”. Sharma denotes upasarga by “preverbs”. There is not really the explicit word for ‘after’ in the sutra, but it is to be understood by the use of the fifth case ending in –ebhyo (-ebhyah). One way to rationalise this is to imagine that the subsequent ap is ‘emanating from’ the end of the word in –ebhyah. So the replacement will be made in the starting (ādih) segment (aL) of the farther or later word (parasya), i.e. dvi+ap = dvi+ īp, giving dvīpam etc. 

The other example

āsīnah ‘seated’

is explained from 7.2.83 īdāsah or īT āsah, “long ī is the substitute of an when the latter comes after the root ās” (Vasu, p.41). Hence, ās+an= ās+īn= ās+īn. Here, since ās is the PRECEDING word, we will have to interpret āsah as being in the fifth case and hence the CONTEXT; but because it is in singular (5/1), the form itself is identical with the sixth case, so that it could as well have been the  substituendum (sthānin)! Such are the pitfalls of this ascetically spare discipline!

Examples of 1.1.55  anekāl śit sarvasya (şaşţhī sthāneyogā from #49)

Again, by #52, the substitution should only be of the last element (alo’antyasya), but here are yet other exceptions: if the substitute (ādeśa) is multi-letter. or if it is a single or multi-letter substitute marked by Ś (ŚiT) , it will replace the ENTIRE original substituendum (word in the sixth case), not just its final letter. Vasu (p.42) cites sutra 2.4.53 bruvo vacih, “in the place of brū let there be vac”, and since there are three letters (Vasu) in v-a-c, it is a total replacement for the whole root brū, giving forms like vaktā, ‘he will speak’. In counting the letters, any boundary-marker is to be ignored.  Another set of examples is

Bhavitā ‘that which is to be, become’
bhavitavyam ‘ought to be, become’
bhavitum ‘for being, becoming’

which presumably emanate from 2.4.52 Aster bhūh, in place of ast- words replace words derived from bhu-, ‘to be’.

An  example for by Vasu for substitute marked by Ś (ŚiT) is 5.3.3. “In the place of idam there is iŚ”, which being marked by the element Ś, is therefore to replace the whole word idam, not just the final letter; even though it is not a multi-letter entity, since it has only one letter i (the marker Ś not being included in the count).

25 More on substitution

A few successive sutras further  expand on the substitution effects. We have

1.1.50 sthāne’ntartamah

Sthāne (7/1) antartamah (1/1)

The vŗtti paraphrase is as follows:

Sthāne (7/1, in the place) prāpyamānānām (6/3, ? of the available) antaratama (closest) ādeśa (substitute) bhavati (is) sadŗśatamah (the most alike)

“A substitute (ādeśa) which is to replace a substituendum (Sthānin) must be most similar to the substituendum” (Sharma).

“When a common term is obtained as a substitute, the likest of its significates to that in the place of which it comes, is the actual substitute” (Vasu).

Of the two translations of the vŗtti, it is rather surprising that Vasu’s is the more cryptic (usually it’s the other way round). Let’s first use Sharma’s more concise terms: the thing which is to be replaced ('for X') is termed the substituendum, or Sthānin (in common language, ‘that which is in place’). The thing which replaces, the substitute ('put Y'), is the ādeśa (in common language, the ‘directive’). The explanation suggests that in various sutras, one may be given a list of substituendums  (Sthānin) and a list of substitutes (ādeśa), without any direction how to match the two lists. In such a situation one matches the pairs according to their similarity.  The word antaratama is a bit puzzling, because I thought antara is actuall a ‘gap, difference’, and tamah is a superlative degree; but I guess one should interpret it as the ‘most subtle of differences’ rather than as the ‘largest of differences’. Of course, the previous sutra, 1.1.49 şaşţhī sthāneyogā is to be carried in the mind as it gives the context to sthāne: the genitive case ending denotes the substitution (in place of) relationship, and the substitute (replacement) that is to be put in place (sthāne) from the available items will be that showing the closest similarity (antaratamah).

Thus far thus good. Now 'similarity’ is said to be of the following types: place of articulation (sthāna), signification (artha), sound quality (guņa), or duration or quantity (pramāņa). Usually we use the place of articulation as the guide (note here the dual meaning of sthāne: ‘in place [of the substituendum]’ or ‘in place of articulation’. A question is raised why the word sthāne is not taken over from the previous sutra by anuvŗtti (ellipsis). As per Sharma, the Kāśikā’s explanation is that the author wanted to emphasise that when there is a choice of criteria, “a similarity based on sthāna ‘place of articulation’ should be considered decisive”. This implies that the word sthāna is used in the sense of ‘place of articulation’ in 1.1.50, which is not the sense used in 1.1.49; thus there would be no possibility of letting the word sthāne be taken as understood in 1.1.50 by anuvŗtti from 1.1.49. However, to me this sounds too limiting, and probably not what the author meant: consider that the basis of judging similarity may be any of the four alternatives given above. I would prefer to understand sthāne as ‘in place’, in both the sutras.

Here’s one more development of substitution;

1.1.51 uraņraparah

Taking out the sandhi,

Uh (6/1) aņ (1/1) ra-parah (1/1)

To be understood by anuvŗtti from the previous sutra #49 is   sthāne.

The vŗtti is as follows:

Uh (of ŗ) sthāne (in place) aņ (aŅ, that is the short and long vowels a, i, u) prasajyamāna (?) eva raparo veditavyah.

“If  aŅ (a, i, u) comes as a substitute in place of ŗ, it is automatically followed by r” (Sharma).

“When a letter of aņ pratyāhāra comes as a substitute for ŗ, it is always followed by a r” (Vasu).

Let’s first deal with the mystery of where the ŗ turned up, when the sutra starts with uh. This is one of the problems of the sutras, that they apply the rules of transformation (noun declension, in this instance) even before we’ve introduced the rules. Here the genitive (şaşţhī, sixth) case of ŗ as a letter or sound, is derived as uh obviously on the analogy of other nouns ending in ŗ: we know, for instance, pitŗ ‘father’ is declined pitā, pitarau, pitarah, and so on, to genitive case pituh, pitroh, pitŗ:ņām, etc. (the colon indicates long vowel). On this analogy, ŗ would be declined – ā, arau, arah, etc., genitive uh, roh, ŗ:ņām. This is slightly absurd (if one may dare to use such a word for this hoary vedānga!), and Vasu actually provides a more student-friendly alternative version ŗkārasya, for ‘the genitive of (the sound) ŗ’, in the vŗtti explanation (Vasu, 1891, p.39).   

The next term, aŅ, is the pratyāhāra (see the relevant Page from the TABS on the top) that includes the vowels a, i, u (short and long).  Together, the first two words denote “aŅ (a, i, u) as a substitute in place (sthāne  is understood from 1.1.49) of ŗ” as shown in the  vŗtti paraphrase. We understand that the replacement of ŗ would be occasioned by other rules elsewhere in the grammar, such as 1.1.3 iko guņavŗddhī (see post #03). When such a rule requires replacement of  ŗ by its guņa, say, we choose the “likest” item from the available guņa sounds a, e, o, which according to the “nearness of place” (Vasu, p.40) is a.

The next issue is the term raparah or ra-parah. The first part ra- refers to the r-sound. The second part para means ‘across, beyond’, like pere- or para- in Greek/Latin. Broadly it says that after the substitution of  ŗ (eg. by a in the above citation), there will be added an r sound. Thus the guņa of ŗ will not be just a, but a-r, ar. In some other contexts, the substitute may be i or u, in which cases the actual replacement will be, respectively, ir and ur.

There is an extension of this principle by analogy which implies that replacement of the letter ļ by a, i or u will call for a subsequent l sound to be added, thus giving al, il, ul. Vasu (p.40) puts it thus: “The r in the text has been taken by some to be a pratyāhāra formed by the letter r and a of l^aŅ; and thus it includes the letters r and l. In that case the sūtra would mean  that aŅ substitutes of ŗ and ļ are always followed by r and l respectively. Thus guna of ŗi = ar, of ļi = al.” Personally, this is puzzling, as it appears to me that the boundary-marker of the pratyahara l^aŅ is Ņ and not ^a or a. To include the sounds r and l in one pratyahara, one would have to use the notation r^a Ņ. The first term uh also would have to be changed to include both the letters ŗ and ļ. The normal alternative would be to use the pratyahara ŗK which includes the letters ŗ and ļ followed by the boundary-marker (or iT), K, and its genitive ŗcah. Maybe that is what the sutra is supposed to have meant, but who are we to try and gild the lily!

24 Metarule: sixth case denotes substitution

We skip forward a few sutras to 1.1.49, which we have referred to previously as a metarule (see Post #3, January 2011):

1.1.49 şaşţhī sthāneyogā

The word division is as follows:

şaşţhī (1/1) sthāneyogā (1/1)

Thus both words are in first case (nominative), singular number. The first, şaşţhī, refers to terms in the sutras that are seen to be in the sixth case, which is the possessive or genitive: ‘of something’. The second word, sthāneyogā, tells us how to understand the possessive: i.e. as referring to ‘in place of’, rather than say ownership or proximity or part-whole relationship and so on. The vŗtti or paraphrase is rather involved, and only a sense of it can be got from the translation in Sharma or Vasu:

Iha śāstre (in this canon [the sutras]) yā şaşţhī (that in the sixth case) a-niyatayogā (? Not a fixed relationship) srūyate (? Which is heard), sā (that [sixth case])   sthāneyogā eva bhavati (is only the sthāneyogā), na anyayogā (not another relationship) sthāneyoganimittabhūte sati  sā pratipattavyā (?).

Formally,

"The force of the genitive case in a sutra is that of the phrase in the place of  when no special rules qualify the sense of the genitive“ (Vasu)

 “A genitive ending (which is not otherwise interpretable in its context) signifies the relation in place of” (Sharma).

The main point here is that the genitive case ending is interpreted in the sense of sthāneyogā, “in place of”  relationship. Thus, in the previous post, we discussed

ik 1/1 yaņah 1/1 (6/1?) samprasāraņam 1/1

where the second word is in the sixth case. So there it was interpreted as the sthāneyogā relation ‘in place of (y^aŅ, the semi-vowels)’, rather than, say, svaswāmi owner-owned  *‘of the y^aŅ’. Put together, it denotes ‘iK (short vowels) in the place of y^aŅ (the semi-vowels)’. The caveat is that the genitive or possessive (sixth) case should not be amenable to interpretation in a normal sense (i.e. other than this technical relation of sthāneyogā ‘in place of’) in the particular context.

Vasu (1891, p.36-37) seems to have somewhat more helpful explanations for this sutra. He suggests that the word  sthāna here has the sense of prasaŋga,  ‘occasion’, i.e. in the occasion of X, then Y, which then translates as ‘in place of’. In the sutra

2.4.52 Aster bhūh, the first word (removing sandhi), asteh,  is the possessive (sixth) case of asti; and the sense of the phrase would be, ‘in place of ast, (use) bhu’ as in forms such as bhavitā (future – will be), bhavitum (gerund – to be), bhavitavyam (passive participle – that has been).  

  

Vasu also unpacks the compound  sthāneyogā as a bahuvrīhi compound, where the first element sthāne is in seventh case (locative): that which has the relationship (yogā) denoted by sthāne (in the place).

23 Substitution of semi-vowels by short vowels: samprasāraņam

Sutra 1.1.45 denotes by the technical term (samjñā), samprasāraņam, the replacement of the ‘semi-vowels’ y, v, r, l (see Coulson, p.14) by the ‘short’ vowels i, u, ŗ, ļ:

1.1.45 igyaņah samprasāraņam

The word parsing (according to Sharma, Vol.II, p.44) is as follows, but I think the case-ending of the second word has been mistakenly shown as 1/1 (nominative case, singular number), whereas in fact it should be 6/1 (possessive case, singular):

ik 1/1 yaņah 1/1 (6/1?) samprasāraņam 1/1

As to the wider meaning of the term, I am unable to come up with anything plausible. However, I do feel the effort should be made, because it may throw light on the grammarian’s thought process and intentions. If I do come across some allusion as to the common meaning of the term, I will insert a note here later.

The explanation of the sutra from the Kāsikā is the following vŗtti:

ik yo yaņah sthāne [bhūto bhāvī vā] tasya  samprasāraņam iti eşā  samjñā bhavati

(I have removed the sandhis in the original, i.e. igyo and ityeşā). This explains that the term (samjñā) i.e. samprasāraņam, denotes the sounds included in the range iK of the Shiva-sutra (Coulson calls these the ‘short’ vowels i, u, ŗ, ļ), which are put in place (sthāne) of the yŅ sounds (yaņah, 6/1) of the Shiva-sutra (these are the ‘semi-vowels’ y, v, r, l - see Coulson, p.14). The tasya means ‘its’; the samprasāraņam OF the yŅ sounds. We may remind ourselves that these upper-case letters in the range-names like the K in iK and the Ņ in yŅ are just artificial boundary-markers, or iT, to mark off the stretch of letters (or sounds) as laid out in the Shiva-sutra  - see the corresponding Shiva-sutra PAGE above! The range-names are akin to our usage of short forms like ‘A to Z’ which literally translates as aL of the Shiva-sutra (all the sounds from a to h, plus L the last boundary-marker). It is interesting that the sound a (alef) leads the rest in both notations!

Two points to note: one is the technical use of the possessive or genitive case (6) in  yaņah. This is part of the metalanguage used in the grammar, as already mentioned in Post #3 (January 2011!) and which will be discussed again when we deal with sutra 1.1.49 shortly. The genitive case is interpreted as ‘in place of’, in the phrase: yaņah, of yŅ, sthāne, in place, hence ‘in place of the  yŅ’. The sutra, however, has the brief version yaņah; the sthāne is sort of ‘understood’!

The second point here is the words [bhūto bhāvī vā] from the vŗtti (paraphrase), which I have put within square brackets (not in the original!). This is actually a subtle point of order which Sharma, for one, makes much of. The point is that the implicated replacement may have already taken place (bhūta, become, past), or (vā) it may be just contemplated or going to happen (bhāvī). If it is the first situation, fait accompli, bhūta, then it can be understood that the replacement sounds iK are referred to as the corresponding samprasāraņam. It is the second situation, imminent or bhāvī, that causes logical nightmares. Sharma explains it thus:

“Normally, one would expect the nominatum (samjñin) to exist at the time when the name (samjñā) is assigned. The difficulty is this: samprasāraņa cannot be assigned unless yŅ is replaced by iK, and yŅ cannot be replaced unless samprasāraņa is assigned. … This is the reason why … samprasāraņa must be interpreted as a name for both bhūta and bhāvī replacements of yŅ by iK.” However, “vārttika three on 1.1.45 offers another possible interpretation. That is, the rule itself provides that both the vowels that substitute for semi-vowels and the process of substituting vowels for semi-vowels may be called samprasāraņa, as is made clear by the different examples cited” (Sharma, VolII, p.45). This is termed the problem of interdependence or anyonyāśrayatva (Sharma, Vol.II, p.45).

My own reaction is that this is splitting hairs somewhat. It’s like ruling out a statement like “There’s no cat here” or a statement “A griffin doesn’t exist”, on the ground that you can’t use the name (samjñā) unless the thing (samjñin) exists. Obviously, if the replacement has already taken place, the term  samprasāraņa would be understood as referring to the replacements; if we are discussing a hypothetical or future situation, it is (or rather, will be!) the process that will be called samprasāraņa. Of course, we could also assume that the term samprasāraņa would refer to those contemplated replacements as they would transpire, but the grammarians seem to have been disturbed by this sort of loose thinking!

It would be instructive to just glance at Vasu’s treatment, and for the really tough ones to look at the Patañjali Mahābhāşya (edition of Subrahmanya Sastri, Annamalai University – see Resources PAGE!). Vasu (Vol.I, p.34) applies the term samprasāraņa “properly” as the name of the iK vowel sound that has replaced the semi-vowel yŅ, but the term is also used “to designate the whole process of the change of semi-vowels into vowels as in VI.1.13; VI.4.131”. Vasu does not expand further on the ‘before-and-after’ conundrum, apparently satisfied that allowing the application of the term  samprasāraņa to either the process or the end-result will suffice for the average student of the grammar. But then, Sharma does feel that Vasu tends to gloss over the niceties (see Resources PAGE).

Patañjali’s treatment is explained in the first chapter of Subrahmanya Sastri’s Lectures, Vol.III. Amazingly, Sastri’s treatment of 1.1.45 only stretches to short of 7 pages, and it dives straightaway into the discussion of what the term samprasāraņa refers to: either to the very sentence yaŅah ig bhavati, or to the resulting letters iK. The same problem of interdependence or itaretarāśrayam (Sastri, Vol.III, p.3) crops up. Since the sutras use different case forms of the word samprasāraņa, the commentator feels it would be reasonable to infer that the term (samjñā) refers to the letters (varņa, p.3). However, in another place the sutras talk about the samprasāraņa of a sentence (vākya) as well. The bhāşyakāra, Patañjali, therefore concludes that the term is given to that which is to come later (p.6). An analogy is given to the sentence, “Spin a cloth of this yarn” (asya sūtrasya śāţakam vaya iti). This has the same problem of interdependence, because “if it is already a cloth, there is no need to spin; if it is to be spun, it is not already  a cloth; it is contradictory to say ‘Spin a cloth’. But he has said ‘Spin a cloth’ anticipating the name cloth; thence I think it should be spun and it will get the name of cloth after the yarn is spun” (p.7). Is it not amazing what a fine yarn these hoary logicians could spin with their meagre threads (sūtra)!

Finally, an example or two should be cited for the feel of this type of substitution; Sharma gives the pairs

Yaj – işţam ‘that which has gone through the sacrifice; desired’
Vap – uptam ‘sown’
Grah – gŗhītam ‘grasped’

To which we can add

Svap = suptam ‘slept’

These examples of past participles (nişţhā, see Post #22, April 2015) are formed from other strings of rules, such as VI.1.15, which can be referred to in Sharma (Vol.II, p.45).

References

(see also the Resources page for downloadable versions, by clicking on Tab on top!)

Coulson, Michael. 1976. Sanskrit. An Introduction to the Classical Language. Teach Yourself Books. Hodder & Stoughton. (many printings, probably new editions).

Rama Nath Sharma, The Ashtadhyayi of Panini, published in 1987 (first edition) and revised and enlarged (second edition) in 2002, by Munshiram Manoharlal Publishers, New Delhi 

Sastri, Subrahmanya. Lectures on Patañjali’s Mahābhāşya, Vol.III (Āhnikas 7 to 9). Published by the author, 1955, Tiruchrapalli.

Vasu, Srisa Chandra. Ashtadhyayi of Panini original 1891-1898, Indian Press, Allahabad. Reissued, in 2 volumes, by Motilal Banarsidass Publishers, Delhi from 1962 (2009 reprint).