Definition: preposition, prefix. The word उसपर्ग originally meant only 'a prefixed word': cf. सोपसर्गेषु नामसु R. Pr. XVI. 38. The word became technically applied by ancient Sanskrit Gratmmarians to the words प्र, परा, अप, सम् etc. which are always used along with a verb or a verbal derivative or a noun showing a verbal activity; cf. उपसर्गाः क्रियायोगे P. I. 4.59. 'These prefixes are necessariiy compounded with the following word unless the latter is a verbal form; cf. कुगतिप्रादयः P.II. 2.18. Although they are not compounded with a verbal form, these prepositions are used in juxtaposition with it; sometimes they are found detached from the verbal form even with the intervention of one word or more. The prefixes are instrumental in changing the meaning of the root. Some scholars like Śākaṭāyana hold the view that separated from the roots, prefixes do not express any specific sense as ordinary words express, while scholars like Gārgya hold the view that prefixes do express a sense e. g. प्र means beginning or प्रारम्भ; cf. न निर्बद्धा उपसर्गा अर्थान्निराहुरिति शाकटायनः । नामाख्यातयोस्तु कर्मोपसंयोगद्योतका भवन्ति । उच्चावचाः पदार्था भवन्तीति गार्ग्यः । तद्य एषु पदार्थः प्राहुरिमं तं नामाख्यातयोरर्थविकरणम् Nir. I. 8. It is doubtful, however, which view Pāṇini himself held. In his Ātmanepada topic, he has mentioned some specific roots as possessing some specific senses when preceded by some specific prefixes (see P. I. 3.20, 24, 25, 40, 4l, 46, 52, 56, etc.), which implies possibly that roots themselves possess various senses, while prefixes are simply instrumental in indicating or showing them. On the other hand, in the topic of the Karmapravacanīyas,the same words प्र, परा etc. which, however, are not termed Upasargas for the time being, although they are called Nipātas, are actually assigned some specific senses by Pāṇini. The Vārttikakāra has defined उपसर्ग as क्रियाविशेषक उपसर्गः P. I. 3.I. Vārt 7, leaving it doubtful whether the उपसर्ग or prefix possesses an independent sense which modifies the sense of the root, or without possessing any independent sense, it shows only the modified sense of the root which also is possessed by the root. Bhartṛhari, Kaiyaṭa and their followers including Nāgeśa have emphatically given the view that not only prefixes but Nipātas, which include प्र, परा and others as Upasargas as well as Karmapravacanīyas, do not denote any sense, but they indicate it; they are in fact द्योतक and not वाचक. For details see Nir. I. 3, Vākyapadīya II. 190, Mahābhāṣya on I. 3.1. Vārt. 7 and Pradīpa and Uddyota thereon. The Ṛk Prātiśākhya has discussed the question in XII. 6-9 where, as explained by the commentator, it is stated that prefixes express a sense along with roots or nouns to which they are attached. It is not clear whether they convey the sense by denotation or indication, the words वाचक in stanza 6 and विशेषकृत् in stanza 8 being in favour of the former and the latter views respectively; cf उपसर्गा विंशतिरर्थवाचकाः सहेतराभ्यामितरे निपाताः; क्रियावाचकभाख्यातमुपसर्गो विशेषकृत्, सत्त्वाभिधायकं नाम निपातः पादपूरणः R. Pr. XII. st. 6 and 8. For the list of upasargas see R. Pr. XII. 6, T. Pr. I. 15, V. Pr. VI.24, and S. K. on P. I.4.60.