Rule. The last word of the sūtra is read as two words, a and anakṣe. After a compound ending in ṛc, pur, ap, dhur or pathin, the affix a becomes the compound's final member (antāvayava). But not after that dhur which belongs to an axle (akṣa). The heading samāsāntāḥ (SK 676) governs, and "ending in ṛc …" is the tadanta reading: the compound as a whole is what the affix follows.
Example.
ṛco 'rdham "half of a ṛc" → compound by ardhaṃ napuṃsakam (A 2.2.2), the ardha going first → ardha-ṛc → the affix a is added as last member → ardharca → the compound is masculine by ardharcāḥ puṃsi ca (A 2.4.31) → ardharcaḥ.
viṣṇoḥ pūḥ "Viṣṇu's city" → viṣṇu-pur + a → viṣṇupuram. The word pur is feminine, and paravalliṅgam (A 2.4.26) would make the compound feminine too; that it is neuter is simply from usage. Pur is listed in the sūtra so that the unwanted viṣṇupūḥ is barred.
vimalā āpo yasmin "in which the waters are clear" → bahuvrīhi vimala-ap + a → vimalāpaṃ saraḥ "a clear-watered lake".
rājño dhūḥ "the king's yoke-pole", dhur used for the burden of the kingdom → rāja-dhur + a → rājadhura → feminine by paravalliṅgam (A 2.4.26), with the feminine affix ṭāp → rājadhurā.
sakhyuḥ panthāḥ "a friend's road" → sakhi-pathin + a → the ṭi of pathin, which is in, is deleted before the taddhita by nas taddhite (A 6.4.144) → sakhipath + a → sakhipathaḥ.
ramyaḥ panthā yasya "whose road is pleasant" → bahuvrīhi → ramyapatho deśaḥ "a region with pleasant roads".
Applies when / blocked by. Three things narrow the rule.
- For ṛc, the forms
anṛcaḥ"without ṛcs" andbahvṛcaḥ"with many ṛcs" are made only when the referent is a reciter (a Ṛgvedin). This is a niyama, a restriction. Where the referent is a text, the affix stays away:anṛksāma"a sāman not built on a ṛc",bahvṛk sūktam"a hymn of many ṛcs". - For dhur, no a when the dhur is that of an axle.
akṣadhūḥis the axle-end, the dhur of the rod on whose ends the wheels turn.dṛḍhadhurakṣaḥ"whose axle-pole is firm" (dṛḍhā dhūr yasya) is a bahuvrīhi. Neither takes a. The sūtra's anakṣe is read as "not of the dhur belonging to an axle", and this reading covers both words — the first has akṣa as its prior member, the second denotes an axle; scholars discuss which reading is best, and this is the one Bhaṭṭoji's commentators settle on. - Two later rules bar or loosen this a: after nañ in a tatpuruṣa (SK 956) and, optionally, after nañ before pathin (SK 957).
Watch out. Dhur does not mean a wooden pole in every example: in rājadhurā it is used figuratively for the kingdom. The rule follows the word, not the meaning; only the axle-sense is excluded.
The division [of the sūtra's last word] is 'a' [and] 'anakṣe'. Of a compound ending in ṛk and the rest [ṛk, pur, ap, dhur, pathin], the affix (pratyaya) a becomes the final member (antāvayava); but not of [a compound] ending in that dhur which belongs to an axle (akṣa). ardharcaḥ. anṛcaḥ and bahvṛcaḥ [are formed] only [when the referent is] a reciter (adhyetṛ). Not here: anṛksāma; bahvṛk sūktam. viṣṇoḥ pūḥ → viṣṇupuram — its neuter gender is from usage (loka). vimalāpaṃ saraḥ.
- commentary BM: in the sūtra 'a' is a nominative with its ending elided (luptaprathamānta); 'samāsāntāḥ' (A 5.4.68) is the governing heading (adhikṛta); the compound is qualified by ṛk etc., hence the tadantavidhi 'ṛgādyantasya' — i.e. ending in ṛk, pur, ap, dhur, pathin; 'apratyayaḥ' = the affix consisting of a. Followed in the bracketed expansions.
- commentary TB (BM agrees): the locative 'anakṣe' is used of a related thing (sambandhin) treated as a locus, so the sense is 'not of a compound ending in that dhur which belongs to an axle'; TB argues that if it were read 'not when akṣa is the prior member' the prohibition would fail in dṛḍhadhūrakṣaḥ, and if 'not when the compound denotes an axle' it would fail in akṣadhūḥ, so the stated reading is preferred as covering both. BM adds that 'anakṣe' attaches only to dhur, being impossible with the others. Followed.
- commentary BM on the forms: ardharcaḥ = ṛco 'rdham, compound by 'ardhaṃ napuṃsakam' (A 2.2.2), a as samāsānta, masculine by 'ardharcāḥ puṃsi ca' (A 2.4.31); anṛcaḥ = avidyamānā ṛco yasya, bahvṛcaḥ = bahava ṛco yasya, 'a Ṛgvedin'; 'adhyetary eva' is a restriction (niyama) whose purpose the counterexamples show — anṛksāma = avidyamānā ṛco yasmin (a sāman not built on a ṛk), bahvṛk sūktam = bahavo ṛco yasmin; vimalāpaṃ saraḥ = vimalā āpo yasmin.
- commentary BM reads 'klībatvaṃ lokāt' as the answer to an objection: since pur is feminine and 'paravalliṅgam' (A 2.4.26) would make viṣṇupura feminine, its neuter gender rests on usage. TB adds that pūr is listed in the sūtra, though the form could be got by compounding with pura, in order to bar the unwanted viṣṇupūḥ. Followed.