Named slaves in early Pali

From a quick search, it seems there are only four named slaves or bondservants in the EBTs.

  • dn3:1.16.1: Disā (probably means “foe” = Sanskrit dviṣa)
  • dn3:1.16.2: her son Kaṇha (“black”)
  • mn21:9.4: Kāḷī (“black”; another Kāḷī is said to “look like a crow” thag2.16:1.1)
  • pli-tv-kd8:1.26.6: Kāka (“crow”)

In the Rig Veda, dāsa (“slave, bondservant”) refers to the “dark-wombed” (kṛṣṇayoni, Rig Veda 2.20.7) foes of the Aryan peoples (Rig Veda 10.22.8) who upon defeat were enslaved (Rig Veda 10.62.10). They do not observe the Vedic rites.

Kautilya has a fascinating passage on slavery in ancient India.

It seems there were three ways of becoming a slave:

  • to be born such
  • to be captured in war
  • to indenture oneself to pay off a debt

Kautilya mostly treats the latter case. Obviously, any form of slavery is abhorrent. Nonetheless, the institution as described by him is quite different from what we might recognize as slavery in a modern context. Indeed, Megasthenes—who was the the Greek ambassador to Candragupta and Kautilya’s possible contemporary—said there was no slavery in ancient India. There were many provisions and protections for the slaves; they could be deprived of their money, the women were protected from rape, but if she had a child both mother and child were freed, and so on.

However, Kautilya also begins by distinguishing between the Aryan, who should never be enslaved against their will, while it is permissible for a non-Aryan (mleccha) to indenture their own children. This introduces a racial dimension to slavery, showing that the by-then ancient divide between the Aryans and the previous native peoples of India had not yet dissolved, as is still true today.

It seems inescapable that the named slaves in early Pali are racially distinct. They are non-Aryans descended from the native peoples of India, perhaps Tamils or other tribal groups, and their slavery is closely connected with their dark skin color.


How about these terms:

  • 領群特 (lingqunte) = vasala/vasala-ka (wicked outcaste) = 旃陀羅 (zhantulo) = caṇḍāla (SA 102 = ASA 268 = Sn 1.7 Vasala-sutta).

Cf. p. 375:

Choong Mun-keat, "A comparison of the Pali and Chinese versions of the Brahmana Samyutta , a collection of early Buddhist discourses on the priestly Brahmanas ", Journal of the Royal Asiatic Society, vol. 19, issue 03, July 2009, pp. 371-382.

  • SN 35.132 (= SA 255):

“These shavelings, fake ascetics, menials, black fellows, offspring of Brahmā’s feet

(“ime pana muṇḍakā samaṇakā ibbhā kaṇhā bandhupādāpaccā”)

Well, these probably also have racial connotations originally. But I have not checked what names if any we know.

1 Like

The individual name (in SA 102 = ASA 268 = Sn 1.7 Vasala-sutta) is the Buddha, Siddhartha Gautama. He is regarded as vasala/caṇḍāla at that time by a Brahmin.

Mleccha means a foreigner (milakkha in Pali), i.e. non-native people. Such people were a evidently a small minority in Indo-Aryan society, and an exception. In SN56.62, the Buddha is reported as saying “Evameva kho, bhikkhave, appamattakā te sattā ye majjhimesu janapadesu paccājāyanti; atha kho eteva bahutarā sattā ye paccantimesu janapadesu paccājāyanti aviññātāresu milakkhesu” (“People of the unknown foreign/mleccha lands are greater in number than the people of Madhyadeśa janapadas i.e. the Indo-Aryans”). So the understanding of milakkhas/mlecchas is that of foreigners who were foreign to Indian society and not the native races/peoples living in Indo-Aryan (northern & central) India. Mlecchas are not used in co-eval Sanskrit literature to refer to the natives of northern & central India, irrespective of their ethnic/racial origins. They are used as far as I know only for foreign people of the Indo-Aryan borderlands or people who were visitors from abroad. In the śatapatha-brāhmaṇa it is a term used for persians.

The rules in the Arthaśāstra permitting foreigners to own slaves (according to their own socio-cultural norms, or within their own frontier-area settlements) were the exceptions to the rule (of not normally enslaving people born within the Indo-Aryan society).

That does not follow from what Kauṭilya says. There is nothing suggesting that mlecchas means previous native races. They are called foreigners (mlecchas), which implicitly means that they were rather non-native (foreign) to Madhyadeśa.

Dāsa means someone performing indentured service in return for a loan or some other debt until they are able to repay that debt or other liability (usually as a result of their own volition) rather than heriditary or involuntary slavery.

Not evident to me from their speech. They seem to be speaking in fluent Indo-Aryan from what evidence we have from the Pali canon, and there aren’t Tamil loanwords, Dravidian names or other Dravidian expressions visible (perhaps I’m missing something here). Presumably they were not native speakers of Indo-Aryan, but that is at best a guess, and isn’t evidenced either. So I don’t see with what evidence you come to this inescapable conclusion. The names of the dāsas/dāsis quoted above are mostly imaginary people, for ex. the Kaṇha quoted in DN3 is a back-formation from the fictitious name Kaṇhāyana, the Kālī in MN21 is a character that appears in a parable recounted by the Buddha and is equally fictitious. In the mind of the Buddha, or of the narrator of the Pāli canon (in his time and place), those names might have had been appropriate to use for dāsas or dāsīs - but the colour of the skins of most other dāsas/dāsīs, or their races of origin, cannot be extrapolated from the names of these parable characters.

Indo-Aryan India was not seen as the land of the mlecchas, rather the opposite. Mlecchas were the people from abroad who sometimes migrated/visited India in relatively small numbers and were unfamiliar with Indian social norms. Long term settlers such as Dravidians who spoke fluent Indo-Aryan would be expected to know the socio-cultural norms of Indo-Aryan India, and they wouldnt be normally called mlecchas in that case…

Also on a point of linguistic history, Prof. Manfred Mayrhofer’s recent Sanskrit-German dictionary (Etymologisches Wörterbuch des Altindoarischen - you can see a review of it here) indicates that Vedic (approx Pre-Buddhist Sanskrit) is almost completely free of Dravidian and Prakrit vocabulary while new words found in later classical Sanskrit vocabulary (from after the Buddha’s time), contains large numbers of Dravidian, Prakrit, and other foreign vocabulary. That indicates that Sanskrit in India remained relatively free of external influence until circa the start of the Mauryan/Buddhist period – compared to later periods when such influence began to show prominently in Indo-Aryan literature. So there isn’t the hard data to indicate that the majority people of early Indo-Aryan India were not the Indo-Aryans themselves (but rather Dravidians, Munda and other non-IA ethnicities). Dravidian influence on Sanskrit is for the most part after the Buddha’s time.


The most prominent one I can think of is Khujjuttara. However, it seems she’s not mentioned as a slave, only as a laywomen in the EBTs

You’re right, mleccha is generally used to mean “foreigner”, not specifically native peoples. But it’s a general term, including native and non-native people seen as foreign by the Aryans. The Mahabharata repeatedly refers to the “mlecchas of the coasts”, while military campaigns against mlecchas most likely refer to conflicts within the subcontinent. Some references here:

Jains, Buddhists and Hindus all see Mlecchas as speaking unclear, non-Aryan languages, eg. Jain Sutrakritanga (cf. Vinaya Parajika 1):

As a Mleccha[10] repeats what an Ārya has said, but does not understand the meaning, merely repeating his words, so the ignorant, though pretending to possess knowledge, do not know the truth, just as an uninstructed Mleccha.

There are extensive lists of mlecchas, which are clearly meant to be all-inclusive of non-Aryans (or more specifically, non-Indo-Aryans), wherever they might be from. Eg. the commentary to the Jain Tattvarthasamgraha:

The unevolved in other parts of the world–of labour–are the savage tribes, the ionians, the mountaineers, the foresters and so on. These are ‘karmabhūmija mleccha ’.

Or in more detail, again from a Jain source:

The Mlecchas— Śākas, Yavanas, Śabaras, Barbaras, Kāyas, Muruṇḍas, Uḍras, Goḍras, Patkaṇakas, Arapākas, Hūnas, Romakas, Pārasas, Khasas, Khāsikas, Ḍombilikas, Lakusas, Bhillas, Aṅghras, Bukkasas, Pulindas, Krauñcakas, Bhamararutas, Kuñcas, Cīnas, Vañcukas, Mālavas, Draviḍas, Kulakṣas, Kirātas, Kaikayas, Hayamukhas, Gajamukhas, Turagamukhas, Ajamukhas, Hayakarṇas, Gajakarṇas, and other non-Āryas also are people who do not know even the word ‘dharma

For Kautilya, mlecchas are those who live in land conquered by a king, may be used as spies, who threaten the king in his palace, but whom nonetheless a destitute king may recruit. All of these are clearly referring to people in India, since Indian kings did not launch campaigns outside the subcontinent.

I don’t think this is right. The only passage I can find (3.2.1.[24]) says:

Such was the unintelligible speech which they then uttered,–and he (who speaks thus) is a Mleccha (barbarian). Hence let no Brahman speak barbarous language, since such is the speech of the Asuras. Thus alone he deprives his spiteful enemies of speech; and whosoever knows this, his enemies, being deprived of speech, are undone.

This is said of the asuras, and hence any foe of Aryan culture. The word mleccha probably originated with a word used by the Sumerians, Meluhha, a land which is sometimes identified with the Harappan civilization. But that’s not Persia, and it was long before the SPB, like a thousand years or more.

Interesting detail, the unintelligible speech of the Asura-mlecchas is “He ’lavaḥ! he ’lavaḥ”. The particle he is rare in Pali, but it is found in the scolding of Vaidehika to her maid Kali: he je kāḷī!

So? Slaves are almost invariably named “black”, which is explicitly identified with their skin color. I really fail to see how any other conclusion can be drawn.

Sure, the influence grew over time. The word mleccha itself is a loan word. But according to Witzel, in the Rig Veda:

we find more some three hundred words from one or more unknown language(s),

That’s about 4% of the vocabulary. His study is well worth a look.

One point he makes (p 19), which I agree with, is “the very hesitant acceptance of non-Indo-Aryan words and forms in the high level, poetic language of the RV”. It is normal that such elevated literary texts will resist contamination from other languages. So it’s likely that the spoken language of the Vedic period, or in any period really, had more loan words. The same thing would apply to the Pali.

Anyway, it’s not really relevant to my point. Slaves named in Pali are “black”. So far as it goes, this agrees with the Arthasastra’s admittedly weak correlation between slaves and mlecchas. But that’s really a secondary point.

My main purpose was to highlight some details of the Pali. There is an ongoing debate as to the extent to which slavery in ancient India has a racial dimension. Clearly it was not as racially centered as the most well-known type of slavery to modern people, namely, the trade of African slaves to the Americas. Nonetheless, there are multiple indications that Indian slavery originated with the subjugation of the dāsa or dasyu peoples, who were natives of India. I don’t know whether these data points from early Pali have been noticed in this debate; I suspect not. I just wanted to highlight them, because however you might consider the matter, they are relevant.

And by the way, even when I disagree with you, I always end up learning a lot! So keep disagreeing!


So, the people of dāsa “slave, bondservant” in India are likely both kaṇha “black” and vasala/vasala-ka/caṇḍāla “outcaste”.

The people of “black” in India are also likely “outcaste”.

But the Buddha likely has some dark skin color (but not “slave” and not entirely “black”), because he is also regarded as “outcaste” by a Brahmin at that time (according to SA 102 = ASA 268 = Sn 1.7 Vasala-sutta: Vasala Sutta: Discourse on Outcasts).

Yes, therefore they are foreigners (perhaps in trading settlements along the coasts) and living along the frontiers - who would be a very small minority.

That is consistent with them being foreigners to Indo-Aryan India. The word mleccha is derived from the root mlecch (which means to speak indistinctly or unclearly, or to blabber).

As a Mleccha repeats what an Ārya has said, but does not understand the meaning, merely repeating his words, so the ignorant, though pretending to possess knowledge, do not know the truth, just as an uninstructed Mleccha.

This doesn’t apply to the dāsas and dāsis living in Indo-Āryan India who do speak and understand Indo-Aryan, so they are not the mlecchas that this quote is describing.

That is true, however they were not normally part of Indo-Aryan society/culture and didn’t speak or understand Indo-Aryan - which is why they were considered mlecchas. They also normally lived in separate settlements. They would not have been synonymous to dāsas. The spies may have been mercenaries, but they are the exceptions, not the rule.

It appears that you are talking about India, but I am talking about Indo-Aryan India, the two are not one and the same. Indo-Aryan India (natively called Āryāvarta, within which there was the Madhyadeśa) were within India (the subcontinent). So the Dravidian-majority areas of the subcontinent (i.e. in South India) would have been mleccha-lands to a person living in Indo-Aryan India (roughly northern & central India). But a Dravidian family who were speaking Indo-Aryan and living in Indo-Aryan India for generations would not be called a mleccha unless their speech or pronunciation of Indo-Aryan was particularly bad as to indicate they were recent immigrants.

This is said of the asuras, and hence any foe of Aryan culture.

The language of the asuras here means the language of the asura-worshipping (i.e. Zoroastrian) achaemenid-era Persians i.e. Avestan, Median, Old-Persian etc (i.e. Old-Iranic). The reason why the śatapatha-brāhmaṇa mentions that no Brahmin should speak in that language is because it was possible to speak in Old-Iranic as it was a human-language (not an imaginary language spoken by mythical Asuras) and it was very close to Old-Indic (both being derived from the ancestral Indo-Iranian). But it was foreign to Indo-Aryan India, and it led to phonetic and morphological confusions in Old-Indo-Aryan (such linguistic babbling are given as a reason by the ŚB as to why someone would be defeated in war, as such speech would spread likely confusion among the rank and file). The ŚB says the asuras were defeated because they babbled. This is repeated in multiple rescensions of the śatapatha-brahmaṇa (kāṇva/mādhyandina etc).

The he 'lavaḥ, he 'lavaḥ (or he 'lavo, he 'lavo in another version , or he 'layo he 'layaḥ in yet another version) is Persian for Sanskrit “he-arayaḥ, he-arayaḥ” (“hey enemies”). arayaḥ is plural for ‘ari’ (enemy). Pali also has sing. ari & pl. arayo (with the same meaning as in sanskrit).

Pali (and Sanskrit) “he” (also “hai” in Sanskrit) is simply a vocative particle synonymous with and pronounced exactly the same as English “hey” and is extremely common in speech when calling someone by name to draw their attention. Pāṇini says in sūtra 8.2.85 haiheprayoge haihayoḥ that the interjections ‘hai’ and ‘he’ are used to call people from a distance.

That is simply “Hey, you - Kālī” (in Sanskrit: he + aye + kālī), the sandhi causes the elision of a in aye, and the ye becomes je in Pāli so “he 'je kāḷī”.

Monier Williams’ dictionary entry for aye = ind. a vocative particle, an interjection (of surprise, recollection, fatigue, fear, passion).

I don’t see the logic of connecting the common-noun mleccha in the śatapatha-brāhmaṇam with the proper-noun (name) Meluhha used by the Sumerians nearly 2000 years before that. So there is nothing probable about such an association (in my mind) - and I think this suggestion too comes from either Witzel or Asko Parpola, and in my understanding the suggestion is incorrect.

Let’s see what Mayrhofer’s etymological dictionary says about the word mleccha. He classfies it as a Vedic (i.e. etymologically Indo-European origin) word, i.e. not a loanword from Sumerian or other source. He mentions the latin word ‘blaesus’ (“to lisp”, or “to stammer”) as a possible IE cognate. That also links it to ancient-greek βλαισός (blaisós) - which has the semantic sense “to be distorted”. I think Mayrhofer’s suggestions here are on much more philologically firmer ground than Witzel’s or Parpola’s.

That “invariability” conclusion cannot be drawn from just two names. It needs many more names from multiple buddhist and non-buddhist sources to establish such a pattern and derive inference therefrom. In the vedic corpus, the names of dāsas are clearly not Dravidian, and are also not indicative of their skin colour, for example Namuci, Navavāstva, Pipru, Suṣṇa, Dhuni, Cumuri, Śambara, Balbūtha etc. Namuci is also mentioned in the Pāli canon, but here he is a personfication of Māra, who is antithetical to the Buddha (while in the Vedas, he is antithetical to Śakra/Indra).

Also both Kaṇha (going back all the way to Proto-Indo-European kr̥snós) and Kāḷī are Indo-Aryan names, one being the pali-form of the name of a prominent epic-god Kṛṣṇa, the name of a student (Kṛṣṇa Devakīputra) of the philosopher Ghora-Āṅgirasa mentioned in the Chāndogya-Upaniṣad, the name of a Ṛṣi (Kṛṣṇa Dvaipāyana, the attributed author of the Mahābhārata - who even has a jātaka named after him) and of numerous other Indo-Āryan elites as well. Kālī is also the name of a Hindu goddess (and she is mentioned in the nearly co-eval Kāṭhakagṛhyasūtra along with other vedic gods: “agniṃ somaṃ varuṇaṃ mitram indraṃ bṛhaspatiṃ skandaṃ rudraṃ vātsīputraṃ bhagaṃ bhaganakṣatrāṇi kālīṃ ṣaṣṭhīṃ bhadrakālīṃ pūṣaṇaṃ tvaṣṭāraṃ mahiṣikāṃ ca gandhāhutiṃbhir yajeta”. These names also occur numerous times in the epics etc (and not invariably in a derisive sense).

Being dark in skin colour (or having a name implying darkness) is therefore not ipso-facto = being non-Aryan or mleccha or dāsa/dāsī.

Vedic as a whole was not a heiratic language (the claim of it being a heiratic language applies even less to the language of the Ṛgveda, than it does to Vedic in general). At the point of composition of the Vedas, the language used in their composition was not heiratic. The early-Vedic language got a heiratic status post-facto (after the elapse of many centuries, when the spoken language had already evolved to late-Vedic or Classical Sanskrit). I don’t know how people like Prof. Witzel make such claims, they make no sense to me. Also see

By Ancient Sanskrit we mean the oldest known form of Sanskrit. The simple name 'Sanskrit' generally refers to Classical Sanskrit, which is a later, fixed form that follows rules laid down by a grammarian around 400 BC. Like Latin in the Middle Ages, Classical Sanskrit was a scholarly *lingua franca* which had to be studied and mastered. Ancient Sanskrit was very different. It was a natural, vernacular language, and has come down to us in a remarkable and extensive body of poetry. (From: Ancient Sanskrit)

So Ancient (Vedic) Sanskrit, as we find it attested in the surviving pre-Buddhist texts, was very much a natural vernacular everyday-language back in the day when those texts were being composed (which is the polar opposite of “heiratic”). It was an evolving language that had multiple dialects - and all this is evident in the Vedic texts belonging to multiple centuries and geographies within Indo-Aryan India.

Even late-Vedic (or Classical) Sanskrit was reduced to the status of a scholarly-lingua-franca gradually many centuries after Pāṇini’s time, but in his own time it was the widely spoken natural language in which most of the Indo-Aryan texts of his time were being written. So Witzel’s opinion about the nature & identity of early-Vedic (attested in the Vedic corpus) is the opposite of the facts.

That is in his own imagination (or rather the imagination of his mentor? - F.B.J. Kuiper). I am not aware what the basis is of their coming to the conclusion that the language of the RV at the point of composition was heiratic, or that there was a very hesitant acceptance of non-IA vocab therein. Have they reasoned it out somewhere?

He thinks there are 300 non-IE words in Ṛgvedic but he is including proper names in them (exhibiting non-standard phonetic properties for an Indo-Iranian IE language, based on comparative philology). Many of these supposedly non-IE words are only found in the Ṛgveda (and have not been found in any other foreign language family), a few are linked to Burushaski spoken by about 0.001% of the subcontinent’s population, which shows that whatever their etymological source (and/or subsequent phonetic distortion), they are not all prominent survivals in later Sanskrit, or in younger IA (or non-IA) languages. Other equally eminent scholars such as Paul Thieme & Mayrhofer dont accept Witzel’s or Kuiper’s propositions about those words. So he is trying to build a generalized theoretical edifice using fringe sources and doubtful etymology.

Certainly not, the dāsas and the dasyus were originally small groups (or tribes) of people from somewhere between India and Iran, perhaps Bactria (to the north-west of the subcontinent), not Indians at all. They are also mentioned in Iranic texts as dāha/dāhae (Iranic for dāsa) & dahyu (Iranic for dasyu) - and they appear to be inimical to both the Indo-Aryans and Irano-Aryans - thus pointing to a strife that presumably goes back to Indo-Iranian times (pre 1500 BCE). In later-Vedic, the term dāsa gets a change of meaning to mean ‘(indentured) servant’ and dasyu too gets a more-generic meaning of “thief”. They would not both appear prominently in Old-Iranic texts if they were exclusively Indian tribes residing within the subcontinent. Robbing appears to be one of the activities the original Dasyus apparently engaged in usually in Vedic times. In later Sanskrit (and Pāli), dāsa is not used to describe any specific tribe, group or ethnicity and being a dāsa (normally used in the singular) does not carry any connotation of foreignness, unlike the word mleccha which is normally used in the sense of a foreign community/group i.e. in the plural (or an individual belonging-to or arriving-from such a group).


Vṛṣala does not mean outcaste. It is semantically close to the name of the servile class (i.e. Śūdra) but is not identical to that class. Vṛṣalas were not seen as a specific class or caste, it is a term used to express social contempt towards individuals or generally indicative of their low-social-rank within Indo-Aryan society but they were not considered uncivilized.

Caṇḍāla (derived from the word caṇḍa “brute/barbaric”) meaning uncivilized/brutish people (towards whom no civilized social contact was possible or desired i.e. effectively untouchable - or with whom contact was forbidden. Outcastes were sometimes equated with them, but the original caṇḍālas (non-outcastes) were evidently forest-dwelling ‘savage’ types. They were described as śvapacas (people who ate dog-meat) or cannibals, or people feeding on corpses etc. Basically people who were most repulsive to Indo-Aryan society, and to brahmins in particular.

Dāsas could be dark (kṛṣṇa) skinned but they could not have been vṛṣalas (who were free people though of low social status) and not caṇḍālas either (who were not accepted as civilized in the first place and who therefore could not be dāsas).

Outcastes were none of these i.e. they were not generally dāsas or vṛṣalas (or a specific caste or class of their own). Out-caste refers to individuals who were formally excommunicated from their former-class and simultaneously banished from habitable society (as a result of seriously breaching moral/social norms or doing serious crimes) and who therefore were reduced to the social status of caṇḍālas (and with whom social contact was not possible or forbidden due to their perversion from ethical/social/moral standards). However those who were excommunicated & banished sometimes settled in other janapadas and didn’t live like caṇḍālas despite being banished from their former society/region. These are comparable to the disrobing/expulsion of monks in buddhism (after they commit a crime that is serious enough as to be seen incompatible with monkhood).

These are social nuances that are not always entirely inferrable from just a knowledge of the Pali canon, and are frequently misinterpreted in non-Indic translations (by people who don’t read Sanskrit texts more generally, and who dont have a rounded picture of early/classical Indo-Aryan society).


So, how do you translate the Pali terms, vasala/vasala-ka, caṇḍāla into English?

How about: “wicked outcaste”?

Caṇḍāla = Barbarians
Vṛṣala (Pāli: vasala) = Low-born / menial / servile

There is nothing wicked or outcaste semantically about vṛṣala whatsoever (an outcaste is by its very definition not a vṛṣala - as the vṛṣala would belong to a jāti and/or varṇa). They were part of Indo-Aryan society, and were commoners, not particularly intellectually distinguished, elite or powerful or wealthy enough as the elite classes.

There is nothing wicked or outcaste about caṇḍāla either prima-facie (but outcastes were socially equated with caṇḍalas sometimes, though the words outcaste & caṇḍāla are’nt synonymous to begin with, and mean different things). Caṇḍalas did not observe the hygiene/purity/ethical/cultural/civilizational norms of the Indo-Aryans. That doesn’t mean they were wicked or out-castes, they were however different, and the Indo-Aryans in general didn’t like their way of life and didn’t let them reside in their villages & cities (along with the rest of the civilized society).


Pali-English dictionary

[«previous (V) next»] — Vasala in Pali glossary

Source: BuddhaSasana: Concise Pali-English Dictionary

vasala : (m.) an outcast; a person of low birth.

Source: Sutta: The Pali Text Society’s Pali-English Dictionary

Vasala, (Vedic vṛṣala, Dimin. of vṛṣan, lit. “little man”) an outcaste; a low person, wretch; adj. vile, foul Vin II. 221; Sn. 116—136; J. IV, 388; SnA 183, — f. vasalī outcaste, wretched woman S. I, 160; J. IV, 121, 375; DhA. I, 189; III, 119; IV, 162; VvA. 260.

Pali-English dictionary

[«previous (C) next»] — Candala in Pali glossary

Source: BuddhaSasana: Concise Pali-English Dictionary

caṇḍāla : (m.) an outcaste or untouchable.

Source: Sutta: The Pali Text Society’s Pali-English Dictionary

  1. Caṇḍāla, 2 (nt.) a kind of amusement or trick D. I, 6≈(=ayogulakīḷā play with an iron ball DA. I, 84). (Page 260)

  2. Caṇḍāla, 1 (Vedic caṇḍāla) a man of a certain low tribe, one of the low classes, an outcaste; grouped with others under nīcā kulā (low born clans) as caṇḍālā nesādā veṇā rathakārā pukkusā at A. I, 107=II. 85=Pug. 51. As caṇḍāla-pukkusā with the four recognized grades of society (see jāti & khattiya) at A. I, 162.—Vin. IV, 6; M. II, 152; S. V, 168 sq. (°vaṃsa); A. III, 214, 228 (brāhmaṇa°); IV, 376; J. IV, 303; PvA. 175; Miln. 200.—f. caṇḍālī A. III, 226; Pv III, 113; DhA. II, 25. See also pukkusa. (Page 260)

The dictionary entry youve quoted above contradicts itself by giving 3 meanings that are mutually incompatible with one other.

A low class cannot (as a whole) be considered outcaste, because all the classes of society are part of the society and are by definition not outcaste. Outcastes are by definition outside the social order. A caste is not a class, and a class is not a tribe. They are not synonymous with each other and while there is a term called outcaste, there is no corresponding outclass or outtribe. So the dictionary entry above is not precise enough. That is why I explained what each word means in my posts above to make it clearer.

Etymologically as I mentioned above, the word caṇḍāla is derived from caṇḍa + ālaC pratyaya (as per sanskrit grammar) which gives the meanings ‘One who is innately fierce/barbaric/cruel’ etc. I dont see how it can mean outcaste. Outcastes were reduced to the status of (and compared to) caṇḍālas, true (and caṇḍālas were socially shunned like outcastes were, so there is an attempt to equate the two identities), but they have different senses, and caṇḍāla by itself could not have primarily meant outcaste.

I already have access to nearly 30 dictionaries - I’d be wary of taking mutually incompatible dictionary suggestions uncritically.

Different dictionaries work differently. A dictionary like Monier Williams simply throws up all the possible meanings that a word is found attested in all literature of any period, with no etymological or grammatical analysis indicated (and some of those meanings are plain wrong). Relying on such a dictionary will be hopeless for anything other than getting a broad and generic idea of all the word’s senses. If you want precision and clarity there is no way other than incisive grammatical understanding, diachronic analysis and contextual interpretation. Also you have to look at the frequency with which a word is used with a particular meaning in literature. That is likely to be close to its primary meaning.


Outcasts can refer to social outcasts. That is, an outcast is someone who is not accepted by a group of people or by society. This is not “outside the social order”.

Vasala, caṇḍāla are certainly part of Indo-Aryan society in the system that they are not accepted by a particular group of people or society.

But it does not mean vasala, caṇḍāla are in fact culturally barbarians.

It appears you are trying to apply/superimpose your generic understanding of the English language term ‘outcaste’ on Pali and Sanskrit vocabulary, rather than trying to find out what those source-words really mean in Pali and Sanskrit.

The range of meanings in which the term outcaste is used in English are not identical to the range of meanings in which the words caṇḍāla and vṛṣala are used in Pali/Sanskrit.

Outcaste in English can mean one or two or three or 4 different things, and vṛṣala and caṇḍāla can mean one, or two or three different other things.

It is not possible to equate each meaning of vṛṣala and caṇḍāla with each meaning of outcaste, or to argue that they must mean exactly the same simply because they are in the dictionary. Dictionary entries are almost always semantic approximations, and the less related the source and target languages, the more fuzzy (less precise) the meanings are likely to be in general in the target language.

There is almost never a complete semantic overlap possible for any Pali and English word i.e. extremely few Pāli words (and their dictionary equivalents) mean the same as each other in all senses in which they may be used.

I have an independent emic i.e. native (or near-native) idea of what they mean, regardless of what a particular dictionary translates the words into, and I can validate that understanding using etymology, grammar and frequency distribution of that word (and its semantic range) synchronically and diachronically in literature. There are also earlier and later Indo-Aryan and/or Dravidian languages to consider how those words are used in those related languages. All these need to be considered, and I try to consider some or all of them when I present my understanding.

Also consider the cultural and social contextual background of the ancient Indo-Aryan society, its social structures and how they operated. If you haven’t yet a clear understanding of those things, you are more likely to misunderstand than understand these Pali and Sanskrit terms. When you say they were part of society you have a modern liberal Western catch-all conception of society, not the classical/historical Indian structured and exclusivist conception of it. I can go on and on about all the ways in which these sort of misunderstandings arise in most modern Western readers but I think I have said enough.

So when you simply present dictionary entries (and a-priori assume their correctness), and view them opaquely and uncritically i.e. you dont look critically at how the source-words mean what they mean, I dont see a possibility of reasoned or incisive discussion.

Vasala, caṇḍāla are social outcasts in India. They are certainly not “outside the social order” in Indo-Aryan society.

You are repeating the same things, and my opinion is that what you are saying in the post above is misconceived. I have nothing more to say than what I have already said above, you are free to hold on to your beliefs. :pray:

You are also repeating the same things, you are also free to hold on to your beliefs.

Thomas, you could be wrong, depending upon when and what you are talking about. Westerners commonly make the mistake of describing varna as caste. The caste system developed later. Of the four varna, which are identified in the purusa sukta, the lowest was not allowed to sacrifice. The gayatri was also forbidden to the lowest varna and women. Whether that still exists or not, I don’t know. So, in that sense, they were outside the “social order.”

Whether than meant shudra were born forever shudra (or lower) or whether they were so low they were simply “food” and lived a menial existence of servitude for the duration of their short lives, I don’t know. But it is a mistake to simply equate caste with varna.

The idea would not be perhaps so different from, for instance, a very rigid application of Catholicism, in which excommunication means that you are damned.