(4) Compounds

**Divided words treated as compound words**: *-paks.a-m. -kucik ¯ a¯* (*m.c.* for *-paks.akucik ¯ a¯*; 15.37c); *khan. d. aroha sma´ ¯ san¯ ¯ı ca vidrav¯ı kurukullika¯h.* (*m.c.* (?) for *khan. d. aroha-´ ¯ sma´san¯ ¯ıvidrav¯ı-kurukullika¯h.* ; 15.56cd);<sup>80</sup> *purvottarapa´ ¯ scimo ca daks.in. advare ¯ s.u* (*m.c.* for *purvottarapa´ ¯ scimadaks.in. advares ¯ .u*; 15.141cd), etc.

**Compound words treated as divided words**: *-bhujasya ¯ m.* (*m.c.* for *-bhujam asya ¯ m.* : 15.32a); *cakredam.* (a fixed expression for *cakram idam.* ; 15.73b and 211b), etc. I have treated the following first words as words without case-endings and not as the first member of the compound: *bhumi arci ¯ s.mat¯ı* (for *bhumir arci ¯ s.mat¯ı* and not *bhumi-arci ¯ s.mat¯ı*; 15.119c); *varn. a nan¯ avicitrañ ¯* (for *varn. am. nan¯ avicitrañ ¯* and not *varn. anan¯ avicitrañ ¯* ; 15.118a); *idañ cakra sam. svedajan* (for *idañ cakram. sam. svedajan* and not *idañ cakrasam. svedajan*; 15.175a), etc.

(5) *Sandhi*

*-a***+***e- > -e-*: <sup>81</sup> *dvada´ ¯ sete* (a fixed expression for *dvada´ ¯ saite*; 15.224d).

**Hiatus-filler** *m*: *nat.¯ı -m- akhy ¯ at¯ a¯* (15.57a); *tu -m- elakam* (15.189b), etc.

**Hiatus-fuller** *r*: *-pran¯. es.u -r- ¯ıks. an. at¯* (15.283b).

**Double** *sandhi*: *tasyaiva* (for *tasya eva ¯* ; 15.155d), etc. A peculiar form of the double *sandhi* (?): *caturthya-¯* (for *caturtha a-¯* ; 15.81b).

**Frozen** *sandhi: bahyato punar ¯* (for *bahyatah ¯ . punar*; 15.152a), etc.

**When the initial** *a* **follows the final** *¯ı* **of the previous word, the initial** *a* **is removed** (alternatively, this is a form of **the exchange of <consonant +** *i***/***¯ı* **> and <consonant +** *ya* **>** mentioned earlier.): *bhaginey ¯ ¯ı sya* (for *bhagineyy asya ¯* ; 15. 157b);82 and *s´¯ıtak¯ı \*sevana¯* (a scribal error of *sivana¯*) (for *s´¯ıtaky asivana¯*; 15.198c).

<sup>80</sup> Alternatively, unless it is a scribal erorr, the final *h.* of *kurukullika¯h.* is an example of the "Redundant final *h.* /*m*/*n* at the end of a *pada ¯* (especially an even *pada ¯* )" mentioned earlier.

<sup>81</sup> For this, see also (Kiss 2015, p. 84). This *sandhi* (*-a*+*e- > -e-*) was certainly acknowledged by the authors of the *D. ak¯ ar¯ n. ava*. Chapter 16 of the *D. ak¯ ar¯ n. ava* includes a discourse that encodes the letters constituting the fundamental mantra. In that discourse, the words *-sahasreka- ¯* (for *-sahasraika- ¯* ) are encoded (16.137ab and 142a).

<sup>82</sup> I do not regard *bhaginey ¯ ¯ısya* to be an irregular genitive of *bhaginey ¯ ¯ı*.

**The other examples of the non-application of the rule of external** *sandhi*: *dadyat¯ dvi-* (for *dadyad dvi- ¯* ; 15.94cd); *iti ak¯ a´¯sa-* (for *ity ak¯ a´¯sa-*; 15.123), etc.

*•* **Non-cerebralization of** *n* **and** *s*: *-kramena* (for *-kramen. a*; 15.33c); *-rup¯ ani ¯* (for *-rup¯ a¯n. i*; 15.171a); *catusput.an¯ am¯ .* (for *catus.put.an¯ am¯ .* ; 15.225c), etc.

(6) The other orthographical peculiarities

**Exchange of short and long vowels induced metrically**: *-var¯ ahi ¯* (*m.c.* for *-var¯ ah¯ ¯ı*; 15.45d), etc.

**A short vowel followed by multiple consonants is treated as being light**: 83 *karn. ikas¯ uryagnyopari ¯* / (the second *pada ¯* of *pathya, m.c. ¯* for *karn. ikas¯ ury ¯ agnyupari / ¯* ; 15.31b), in which the fifth syllable is treated as being light.

• **Gemination of consonants after -***r* **and the degemination of** *t* **before** *-r* **and** *-v*.

*•* **Exchange of** *tr.* **and** *tr***i in cardinal and ordinal numbers**: *trit¯ıyañ* (for *tr.t¯ıyañ*; 15.132a), etc.

*•* **Exchange of aspirated and unaspirated sounds, exchange of retroflex and non-retroflex sounds, exchange of voiced and unvoiced sounds, and the exchange of** *s´***,** *s.* **, and** *s*: *sma´sana ¯ m.* (for *sma´ ´ sana ¯ m.* ; 15.3c); *bin. d. ipalakam ¯* (for *bhindipalakam ¯* , 15.36d); *ratnes.ika¯* (for *ratne´sika¯*, 15.74b); *ganika¯* (for *gan. ika¯*, 15.89a); *gat.t.ik¯ı* (for *khat.t.ik¯ı*, 15.89c); *bhibh¯ıs. an. a´s* (for *bibh¯ıs. an. a´s* = *vibh¯ıs. an. a´s*, 15.103c), etc.

**Exchange of** *su* **and** *sva*: <sup>84</sup> *-surupakam ¯* (for *-svarupakam ¯* ; 15.11d); *suc¯ı* (perhaps for *svac¯ı* = *svac ´ ¯ı*, *m.c.* for *svapac ´ ¯ı*; 15.87b), etc.

*•* **Omission of** *y* **in a <consonant +** *ya¯* **or** *y¯ı* **>**: *aks. obh¯ı* (for *aks. obhy¯ı*, 15.74a); *nairatm¯ a¯* (for *nairatmy ¯ a¯*, 15. 76c); *rupy ¯ ar¯ up¯ adi ¯* (for *rupy ¯ ar¯ upy ¯ adi ¯* , 15,285d), etc.

**The other Middle-Indic forms of Sanskrit words (including the peculiar words that can be often found in the old Sanskrit manuscripts of other scriptures belonging to the Sam. vara tradition)**: <sup>85</sup> *pan. n. a-* (for *parn. a-*; 17d); *sani´ ´ scaram.* (for *sanai´ ´ scaram.* ; 15.42c); *kallaval¯ ¯ı* (for *kalyapal¯ ¯ı*; 15.89b); *sena¯* (for *syen ´ a¯*; 15.128c); *daddar¯ı* (for *dardar¯ı*; 15.129b); *-khura-* (for *-ks.ura-*; 15.191d); *vijju* (for *vidyut*; 15.192a); *ucchadayet ¯* (for *utsadayet ¯* ; 15.251d); *urddham ¯ .* (for *urdhvam ¯ .* ; 15.254d),<sup>86</sup> etc.

<sup>83</sup> For this, see also (Hatley 2018a, pp. 2–3).

<sup>84</sup> It often occurs that *sva* is wrongly transcribed as *su* because of their possible similarity in the shape of a letter and pronunciation. However, in the critical edition, I have kept *su* and noted that it is used as *sva* in the critical apparatus if several materials support it.

<sup>85</sup> For those Middle-Indic (Prakrit or Apabhram. sa) forms, I have consulted (Tagare 1948) and the ´ *Paia-Sadda-M ¯ aha ¯ n. n. avo* (Sheth [1963] 1986). By "the old Sanskrit manuscripts of other scriptures belonging to the Sam. vara tradition," I indicate the palm leaf and old paper manuscripts (around the 12th–15th century) of the *Cakrasam. vara* (Oriental Institute in Vadodara 13290), *Abhidhanottara ¯* (IASWR I-100 = NGMPP E1517/7 and Asiatic Society in Kolkata G10759), and *Vajrad. aka ¯* (Matsunami 343 and Asiatic Society G3825).

<sup>86</sup> The word *urddha ¯* was already used in the oldest Sanskrit manuscript of the *Abhidhanottara ¯* copied in the first half of the 12th century (NS 258) (Sugiki 2019, pp. 36, 39), which is possibly not so long from,

**Forms of words that seem peculiar to manuscript A and its transmissional lines**: *pit.t.ani ¯* (perhaps from the verb *pit.t.aya* or noun *pit.a*; 15.40b); *d. aka-¯* (for *d. hakka¯*; 15.41a); *karbhara*- (for *karbura*- 15.124c); *bila¯d.¯ı* (for *bid. al¯ ¯ı*; 15.128a); *pitulasya* (for *pitr.vyasya* and not a corruption of *pitur asya*; 15.157d), etc.

In the edited text, I have indicated the peculiar words with underlines, and in the critical apparatus I have provided instructions on how to read them as follows: "-bh¯ıs.an. a¯" in the edited text and "-bh¯ıs.an. a¯ (for -bh¯ıs.an. am) ]" in the apparatus (15.34b). This means that the grammatically irregular "-bh¯ıs.an. a" is used ¯ for "-bh¯ıs.an. am". In cases of peculiarities induced to accommodate the meter, I have indicated those words with the sign "*m.c*." (metri causa) in the critical apparatus. For example, "-bhujasyam ¯ . " in the main text and "-bhujasyam ¯ . (*m.c.* for -bhujam asyam ¯ . )" in the apparatus (15.32a) mean that the grammatically irregular "-bhujasya ¯ m. " is used for "-bhujam asya ¯ m. " to accommodate the meter. In cases of peculiarities in the external *sandhi*, except for some cases that I consider need indication, I have not indicated them with underlines in the edited text or with their standard *sandhi* forms in the critical apparatus. As for the removal of the initial *a* after the final *¯ı*, I have indicated the removed *a-* with an *avagraha* ("bhaginey ¯ ¯ı 'sya") in both the main text and critical apparatus.

To reduce the risk of emendation, when any word in manuscript A is emended, I have represented all emended letters in bold (e.g., "savya¯**vasavyato**"; 15.34c). However, as for the emendation of the orthographical peculiarities, only the emended letters are represented in bold (e.g., "**sma ´** s´ana-"). As for the gemination of consonants ¯ after -*r* and the degemination of *t* before *-r* and *-v*, only the emended consonants are represented in bold (e.g., "sa**rv**a-" and "sa**ttv**a-"). The letters and the part of a letter that are illegible in manuscript A because of damage to leaf or blurring are also represented in bold (e.g., "**nan¯** a-," which means that I have restored the part ¯ *nan¯* from the other source). By these, one can find and check easily what the word in manuscript A is noted in the critical apparatus.

#### *3.2.2. Metrical Peculiarities*

I consider all verses in the edited chapter as *anus.t.ubh* verses. Among those that are rightly metrical, most verses are *pathya¯*, and the following verses are *vipula¯*:

na-vipula: 15.125ab, 181ab, and 258cd. ¯ bha-vipula: 15.107ab and 157ab. ¯ ma-vipula: 15.76cd, 83ab, 90ab, 238ab, and 240cd. ¯

or almost contemporary, with the date when the extant version of the *D. ak¯ ar¯ n. ava* was compiled (see Section 2 in this monograph). For *urddha ¯* , see also (Goodall 2015, p. 127): "It is possible that *urddha ¯* is transmissional, but the possibility that it is authorial cannot be excluded."

ra-vipula: 15.37cd, 65ab, 141ab, and 236ab. ¯

There are also many verses in the edited chapter in which the metrical rule is not followed strictly. One can find odd *pada ¯* s placed in the one of even *pada ¯* s and vice versa, and hypermetrical *pada ¯* s, hypometrical *pada ¯* s, and the other unmetrical *pada ¯* s in which heavy and light syllables are incorrectly placed. The meter is relatively loose throughout the chapter in appearance. However, for the *pada ¯* s that are unmetrical in appearance, a reciter might have skipped reciting a short syllable (syncopation), added a short syllable,87 lengthened a short vowel or shortened a long vowel (see Schott's idea of "freedom of lengthening or shortening"),88 or recited the syllables rapidly or slowly in pronunciation to accommodate the meter.
