Stages of samādhi in Buddhism and Yoga by L. S. COUSINS
(this html doc derived from pdf scan of print article lannce-cousins)
V I T A K K A / V I T A R K A A N D V I C A R A
The two terms vitakka/vitarka and vicāra are crucial to the understanding
of the stages of samādhi in both the Buddhist tradition and in the influential yoga tradition attributed to Patafijali. However, at present interpretation is often dominated by notions derived from later commentarial sources.
Such notions, although in themselves of great interest, create an artificial
appearance of difference between the two traditions which is probably
unjustified. It suffices to note the marked difference in English renderings of
these two words in translations from Sanskrit and from Pali.
This is exacerbated by the, no doubt inevitable, tendency to treat the
Buddhist and Brahmanical traditions as if they operated in complete
isolation from one another. In fact it is clear that each has both influenced
and been influenced by the other in numerous ways. Buddhist origins are
obviously from a milieu in which both orthodox and heterodox Brahmanical
ideas and practices were ubiquitous. Not surprisingly influences from and
reactions to Vedic traditions pervade the early Buddhist texts. Subsequently,
after Buddhism's rapid growth and early creative period, influences are for
a while mainly, but not exclusively, from Buddhism to Brahmanism. After
the formation of classical Hinduism and during the gradual decline in
importance of Buddhism and Jainism which took place from the Gupta
period onwards, it is clear that Buddhism borrows much more than it
contributes. No doubt this is what one would expect, but it seems surprisingly little recognised.
The present issue is a case in point. Influences from Buddhist sources (to
my mind, very frequent) on the Yoga-sutra are often minimized or ignored. 1
In the particular example with which we are concerned here the Yoga-st~tra
is often seen as having a distinctive analysis of the stages of samādhi. I
think this is a mistake, partly due to focussing on later Buddhist literature
rather than on the canonical account. The reason for this is possibly the fact
that the canonical material often needs to be approached through the early
abhidhamma literature which is less studied than the sutta material.
The most important source for this purpose is the first book of the
Abhidhamma-pitaka, the Dhammasangani. This gives mnemonic registers
for both vitakka and for vicāra. For the nature and function of these
registers I refer the reader to my article: "Pali Oral Literature".2 It is
Indo-Iranian Journal 35: 137--157, 1992.
9 1992 Kluwer Academic Publishers.Printedin the Netherlands.1 3 8 L . S . C O U S I N S
sufficient to note that these registers give us a clear picture as to what these
terms were understood to mean at this time once the suttanta contexts to
which they refer have been examined.
(vitakka in the Dhammasangani)
V I T A K K A IN T H E D H A M M A S A i q G A N I
The dhammuddesa for vitakka in the Dhammasangani is as follows:
1. takka 2. vitakka 3. Saṇkappa 4. Appanā
5. vyAppanā 6. cetaso abhiniropanā 7. Sammā-satikappa
Unusually for the Dhammasangani the complete register for vitakka is
already to be found in a single location in the nikdyas, namely in the
Mahdcattdrfsaka-sutta. 3 This discourse is an abhidhamma-style analysis of
the Eightfold Path. In fact the sutta reads suspiciously as if it were itself
based on the Dhammasangani, but if so it is difficult to explain why no
additional sources can be found for some of the terms. We must then
assume that this sutta is the source of this Dhammasangani register and
presumably of much of the methodology of the Dhammasangani, but it is
surprising that no additions have been made. Perhaps the list was already
too established in the tradition to allow of amendment. It would be interesting to know if the corresponding sutta preserved in Chinese contains the
Taking the terms of the register in order:
This occurs in a number of contexts in the earlier literature, but can always
be rendered by 'speculation'. The more specific later meaning of (systematic) logic would be anachronistic, while the translation sometimes given of
'doubt' is incorrect for the nikAyas. The context which the Dhammasangani
or its source probably has in mind is one which occurs in the Brahmajdlasutta :4
9 some mendicant or brfihmana is speculative (takkin) and inclined to investigation
(vfmam. sin). He says that which is beaten out by speculation, that which is attended by
investigation . . . .
In the Ahguttara-nikdya we find the statement that one should not believe
anything by reason of speculation (takka-hetu). 5 Another important sutta
formula also occurs in the Brahmajdla-sutta: 6
There are, monks, still further truths (dhamma) -- deep, hard to see, hard to comprehend,VITAKKA/VITARKA AND VICARA 139
peaceful, excellent, outside the sphere of speculation (atakkdwacara),subtle, (only) to be
known by the wise -- which the Tath~gata makes known after having himself comprehended
them by his higher knowledge and after having directly experienced them.
Similar passages occur in several contexts concemed with the truth assessment of views or wisdom. 7 Finally in the canonical accounts of the request
of Brahma Sahampati the same set of epithets is applied to the dhamma
which the Buddha has reached. 8
To go by the position of right view as last in the list it would seem that
an ascending order is intended. If so, the implication is probably meant to
be that speculation is a rather weak and inferior form of thinking. Certainly
the commentaries have little difficulty in interpreting these contexts in terms
of their understanding of vitakka as the fixing of the mind on an object of
thought or sense.9 For them speculation is merely a form of weak vitakka
whose object is constantly changing. So the term takka-pariydhata recalls
the commentarial definition of the function (rasa) of vitakka as dhananapariydhanana: "for by means of this the yog6vacara makes the object struck
by vitakka, struck around by vitakka". 1~In the present context it would be
easy to interpret takka-pariydhata as meaning that the speculative complex
of ideas which arises in weak mentality requires application of the mind
from many different angles.
The word vitakka occurs frequently in definitions and explanations of
samādhi or jhdna, but is not explained in that context. Apart from this I
have collected about forty other passages from the four nikdyas; there are
probably some more. It is clear that it can always be rendered as 'thinking'
or 'thought', although it is unlikely that this would have the same significance as the concept does for us today. Of course this is even more unlikely
among a community containing many contemplatives. It may therefore be
the case that thought was already pictured as essentially the activity of
bringing different objects into firm focus before the mind's eye -- be those
objects thoughts or mental pictures. Such a view of the matter would after
all be very natural to people with a very highly developed eidetic faculty.
Apart from the above-mentioned accounts of jhdna and the like, vitakka
occurs most frequently in passages referring to the three skilful thoughts or
the three unskilful thoughts or all six together i.e. thought connected with
desire or with desirelessness, with aversion or with non-aversion and with
cruelty or non-cruelty. H Less commonly it is found as part of a series. 12
(There are of course many similar sequences which do not include vitakka140 L . S . COUSINS
at all.) In a number of places it means simply thought or thinking in a fairly
general sense. 13 A few less usual contexts connected with samādhi can be
added. 14 Also we are told that vitakka and vicāra are the activities which
fashion speech: "when one has thought and examined (vicdretvd), afterwards
one utters speech. ''is There is also one discourse which applies the genre of
riddle and answer to the subject of Saṇkappa-vitakkā. 16It is clear from this
and one other passage that Saṇkappa and vitakka are not always identical
in meaning. 17
This should perhaps mean thought formation rather than thought, but not
surprisingly it does not in practice seem greatly differentiated in its use
from vitakka. For example, in a number of contexts the same division into
three unskilful and three skilful types is found. 18 In a general sense of
'thinking' we find 'remembering thoughts' (sara-Saṇkappa) used a number of
times in ways obviously related to the usage of the three unskilful
thoughts) 9 Several times we have expressions like 'due to that' (e.g. fame
and gain) 'he is happy and his purpose is fulfilled (paripun.n.a-Saṇkappa)'.2~
This appears to be the only context where the translation 'purpose' is
required, although it is a possible alternative in some cases, and may
perhaps be appropriate in some passages where safzkappa and vitakka are
juxtaposed or differentiated.
Finally the use of Saṇkappa as part of a series needs to be mentioned. 2~
This is closely parallel to similar uses of vitakka. It is especially frequent to
juxtapose sa~fifi with either vitakka or satikappa. It is emphasized that
sahfifi arises dependent upon the sense objects and corresponding Saṇkappa
arises dependent upon safihfi, but the converse is not the case. This seems
to mean that only if there is say a visual stimulus (~pa-dhdtu) can there be
recognition of the visual object (i.e. rfipa-safihd); only if a visual object has
been recognised can there be thoughts about what has been seen (r@aSaṇkappa). The precise degree of introspective acuteness envisaged is
This occurs only in the one nikāya context previously mentioned. In
commentarial usage it signifies the absorption accompanying strong concentration. The word may already occur in this sense in a doubtful passage in
the Pet.akopadesa, a treatise which may not be long after the early abhidhamma works in date. 22 In the nikdyas verbal forms of appeti occur only
in the sense of 'to flow into (e.g. the sea)' < apyeti.23There are, however, aV I T A K K A / V I T A R K A A N D V I C A R A 141
number of Vinaya passages where it appears to mean 'to fix' < arpayati.24
The same derivation is implied in Vibhahga passages which use it in the
sense of 'made to go away' i.e. 'removed'Y This is the standard etymology
in the later tradition, both in Pali and in Buddhist Sanskrit, no doubt
rightly. 26 It is perhaps just possible that the meaning of Appanā in the
Dhammasangani register is influenced by the sense of 'flowing into' but on
the whole it seems adequate to take it as meaning 'fixing'.
This also appears only in the Mahdcattddsaka-sutta as far as the nikdyas
are concerned, but vyappita is found in the same Vibhahga context mentioned above in the sense of 'completely gone'. The commentary is no doubt
right to interpret this as either augmentation with a prefix for stylistic
adornment or an intensified form of Appanā.27 So it should probably be
translated as 'firm fixing'.
6. Cetaso abhiniropanā
This too does not appear elsewhere in the nikdyas. It is usually taken as
deriving from abhi + ni + causative of RUH -- literally, 'transferring the
mind onto (an object)'. This is probably correct, but it is worth noting that
BHSD gives a form abhinir~payati, which suggests an alternative derivation
from RfYP. It is this which must be intended by the Netti-pakarana
(abhiniropeti) and the Petakopadesa (niropayitabba) in their explanation of
nirutti.28In the latter case it can be translated: 'should be given this form'.
This would offer an alternative rendering for the Dhammasangani register
of 'mental forming' or 'mental defining'. However, abhiniropanā occurs in
the Patisambhidd-magga as the standard epithet for Sammā-sa~ikappa and
also for v i t a k k a . 29 It is also found in one passage in which it defines the
activity of resultant mind element; here it must refer to the fixing of a sense
object in the mind? ~ Although the date of the Patisambhidd-magga is not
known, it must be earlier than the latest of the canonical abhidhamma
As the second link in the Eightfold Path this has an important place. The
nikdyas define it in exactly the same way as the three skilful vitakkas or
satikappas.31 Likewise micchd-Saṇkappa is defined in exactly the same way
as the corresponding types of skilful thought. So a translation by 'purpose'
can be ruled out -- vitakka cannot mean purpose. Moreover there are
contexts in which such a meaning is hardly possible:142 L.S. COUSINS
But although there really is another world, he has the view that there is no other world --
that is his wrong view; but although there really is another world, he forms the thought
(satikappeti) that there is no other world -- that is his wrong thought (micchd-salikappa);
but although there really is another world, he utters speech to the effect that there is no
other world -- that is his wrong speech.3z
Here the sequence is clear. If one's way of seeing is flawed, then the way in
which one's thought will take form in the understanding will be flawed and
likewise the way in which one expresses that understanding in speech.
VICARA IN THE DHAMMASAIqGANI
The dhammuddesa for vicāra in the Dhammasanganis as follows:
1. cāra 2. vicāra
4. upavicāra 5. cittassa anusandhanatd
Again taking them in order:
This occurs in a few passages in opposition to vihdra -- 'wandering' as
opposed to 'abiding in one place', but such general uses do not seem very
relevant to the Dhammasangani. 33 It is possible that some usage which I
have not been able to identify is referred to here -- perhaps cāra as the
second member of some compound. More probably it is used here simply
to indicate a mobile aspect of thought -- its 'wandering'; this would be
appropriate in opposition to 'fixing' as the chief feature of vitakka.
Literally interpreted, this might mean either 'constant wandering' or 'that
which causes the (mind) to wander in different directions'. In practice it
almost always occurs in conjunction with vitakka, while in the nikdyas
vicdreti is usually found with vitakketi. This is nearly always in contexts
associated with jhdna or samādhi. At least once, however, it is part of a
Although the noun does not occur in the nikdyas, the verb anuvicdreti is
found in a few passages, always preceded by anuvitakketi. This would of
course literally mean 'causing to explore', but it is clear from the Maffhimanikdya passages that the use of these two verbs together is intended merely
to indicate the repeated application of vitakka and vicāra;
anu is here simply a prefix indicating repetition.35
The Ahguttara-nikdya usage is
similar, although it almost always occurs there in the p h r a s e : . . , dhammam
cetasd anuvitakketi anuvicdreti manas~nupekkhati ' . . . applies vitakka and
vicāra with the mind to the dhamma, mentally examines the dhamma. '36
VITAKKA/VITARKA AND VICARA 143
The verb upavicarati (used in close conjunction with the noun) means 'to
frequent'. The noun means that which the mind frequents and hence a
sphere of activity or range of interest. 37 Its inclusion in the Dhammasaligani
register is obviously based upon the formula sometimes referred to as the
eighteen manopavicāra: 38 "After seeing a visible object with the eye one
frequents a visible object which is the basis for pleasant feeling" -- the
number eighteen is reached by utilizing three types of feeling in conjunction
with six senses. This list is found in contexts concerned with the same kind
of process that we find described in stages five to eight of the dependent
origination formula. So it is closely related to the use of vitakka and vicāra
as part of a series.
5. Cittassa anusandhanatd
This may mean either 'explorativeness of mind' or 'a state of constant
uniting of the mind'. The former seems the most likely of the various senses
of the Sanskrit verb, while the latter is the interpretation of the commentary: "it is a state of constant uniting of the mind because it constantly
unites the mind to the object and holds it, just as one joins an arrow to the
bowstring and holds it there.''39 This is not impossible, but in view of the
sixth item of the register, investigation or exploration seems more likely.
Only a verbal form occurs in the nikdyas and only in one doubtful
This means 'careful examination' or 'constant examination'. Anupekkhati
occurs in the nikdyas in two formulae. One was cited above under upavicāra. The other, which is much the more frequent, may be translated:
"dhammas are heard much, remembered, practised aloud, mentally
examined (manas~nupekkhita), well penetrated with insight.''41144 L.S. COUSINS
THE TWO REGISTERS
The two registers may then be translated as follows:
1. speculation 2. thought 3. thought formation 4. fixing
5. firm fixing 6. applying the mind 7. right thought formation.
1. wandering 2. wandering about 3. repeated wandering about
4. frequenting 5. explorativeness of mind 6. constant examination.
In the first case the complete register is derived from a single Ma]]himanikdya passage and three items occur only there. The others are used fairly
widely. For vicāra convincing nikdya contexts exist for at least items 3, 4
and 6 of the register.
THE LATER PALI TRADITION
Vitakka and vicāra occur in a number of passages in the later canonical
literature, but these do not add significantly to our understanding of their
meaning.42 Important information is however to be found in several
paracanonical works and in the commentarial literature. These can be taken
in approximate chronological order:
This is probably the oldest Pali work we have outside the Canon itself. It
shows relatively little influence from the abhidhamma, presumably because
it is in effect a general commentary on suttanta. Vitakka is defined as the
first alighting (of the mind on an object), while vicāra is the exploration
(vicarana) of what has been understood (by vitakka). 43 It goes on to explain
in terms of the initial perception of someone coming in the distance. Vitakka
understands that it is a man or a woman and recognizes colour and shape.
Those thinking (vitakkayanto) further investigate (uttari upaparikkhanti) as
to whether the person is virtuous or otherwise, rich or poor -- this is vicāra.
The next sentence is corrupt, but appears to associate vitakka with fixing
(appeti 44) and vicāra with exploring and conforming (or following).
There follows a simile in which vitakka is compared to the striving of a
bird (on taking flight) while vicāra is compared to the subsequent stretching
out of the wings (in flight) which does not involve so much effort. The
intention appears to be to indicate both the subsequent nature of vicāra andVITAKKA/VITARKA AND VICARA 145
its greater subtlety. Later commentaries specify the difference as between
the trembling of the mind at the time of first arising and a subsequent
Several subsequent passages are corrupt, but some further points are
clear. Another simile is given which contrasts silent recitation with (subsequent) contemplation. In view of what follows the reference is probably to
contemplation of the thirty two parts of the body. The two terms are related
to the four discriminations (patisambhidd) and to the stages in the development of insight knowledge.46 In the latter case at least vicāra is compared to
the higher stages. "For one established in the two, bodily and mental
suffering does not arise; bodily and mental happiness does arise. Mental
happiness produced by vitakka in this way is joy (piti); bodily happiness is
Some additional points and similes are given in the Milindapatiha.48 These
must be quite early as this portion of that work is cited by Buddhaghosa
and others with specific mention of N~gasena. Vitakka is given the characteristic (lakkhana) of fixing (Appanā) and this is explained as similar to a
carpenter fixing a thoroughly prepared piece of wood in a joint. vicāra has
the characteristic of pondering over and over (anumajjana -- literally
repeated rubbing or threshing). It too is illustrated by a simile. "Just as,
O king, a bronze gong, which has been struck, afterwards sounds repeatedly
and echoes on.49 Vitakka, 0 king, should be seen as like the striking; vicāra
should be seen as like the sounding repeatedly."
This work is now extant only in Chinese, but is clearly an important source
of the Visuddhimagga. Its account contains most of the material in the
Petakopadesa passage with the addition of a version of one of the similes
from the Milindapahha and an analysis of vitakka and vicāra in terms of
their characteristic, rasa, paccupatthdna and padatthdna.5~This last is a
standard method of analysing dhammas in the commentarial literature and,
as lqfmamoli has suggested,51 is probably in part derived from the sixteen
hdras of the Petakopadesa.
4. The Works of Buddhaghosa
A detailed treatment of these two terms is found in three of the works
attributed to Buddhaghosa.52 All three are plainly based on a simplification
and tidying up of the Vimuttimagga. It is unlikely that Buddhaghosa had146 L . S . C O U S I N S
direct access to the Petakopadesa; material in his writings derived from that
source is clearly secondhand. In fact even the simile from the Milindapafiha
is normally cited as a bell, just as in the Vimuttimagga, whereas in the
Milinda itself it is a gong.53 The Dhammasangani commentary gives the
bell, but also cites the Milinda directly, either from memory or from a
different version, as there are some variations from the text we have --
most notably the gong is cited as a drum.
Of the three commentaries the Vinaya commentary is fairly close to the
Vimuttimagga version with relatively little additional information but
omitting some of the less comprehensible ideas from the Petakopadesa.
Probably the most important addition is the new simile of the bee --
vitakka is compared to a bee with its mind following a scent that drops
straight onto a lotus while vicāra is compared to the bee's wandering over
the lotus after it has alighted.
The fullest account is given in the Dhammasangani commentary. The
Visuddhimagga gives a rewritten and simplified version of this. Both give a
series of new similes. If someone is firmly gripping a dirty metal vessel with
one hand and rubbing on (anumajjana) powder or oil with a cloth, then
vitakka is like the hand which grips firmly and vicāra is like the hand which
rubs. If a potter who is making a vessel has spun the wheel with the blow of
a stick, then vitakka is like the hand which presses down (to hold the clay)
and vicāra is like the hand which moves about here and there (to shape the
clay). Similarly vitakka applies (the mind) (abhiniropana) just like a fixed
pin which has been fastened in the middle when someone is drawing a
circle. VicSra ponders over (anumajjana) (the object) just like the pin which
circumscribes the circle.
Elsewhere abhiniropana and anumajjana are given as the respective
characteristics of vitakka and vicāra. This must come from an old commentarial passage (giving exegesis of the word Tathdgata) which describes the
characteristics of various dhammas.54 Closely related to this is a rewritten
version of the Petaka's simile of the bird. Vitakka, which "occurs by way of
applying the mind to its object", is compared to the movement of a large
bird flying in the sky which takes the air with both wings and then holds its
wings steady; for vitakka becomes one-pointed and then enters absorption
(appeti). vicāra which "occurs by way of pondering over the object", is
compared to the movement of the bird when it swiftly moves its wings in
order to take the air; for vicāra ponders over the object. The Dhammasahgan,i commentary specifically attributes this simile to the [old] at.t.hakathS.55
The same work is the only one to give another simile -- just as one goes upV I T A K K A / V I T A R K A A N D V I C A R A 147
into (drohati) a royal palace depending upon a friend or relative who is a
courtier, so the mind takes up (drohati) an object in dependence upon
N O R T H I N D I A N B U D D H I S T T E X T S
The Abhidharmako~a-bhdsya 56 defines vitarka as cittauddrikatd 'grossness
of mind' and vicāra as cittasfiksmatd 'subtlety of mind'. The Abhidharm~vatara gives a similar definition57 which must be fairly old, since it appears
also in the Jfidnaprasthdna ~8 -- one of the latest of the Sarvfistivfidin
canonical texts -- and in the Abhidharmahrdaya. 59 This application of the
distinction between gross and subtle does not appear in the Pali tradition
before the Vimuttimagga and is therefore probably of Sarvfistivfidin origin.
Noticeably, whereas the Vimuttimagga probably gave it in the form 'grossness of mind, etc.', Buddhaghosa refers simply to grossness, etc. This is
significant in the light of the discussion in the Abhidharmakoga-bhdsya,
where Vasubandhu criticizes the above definition. As he points out, grossness and subtlety are relational terms (in abhidharma) and should extend as
far as the highest level of existence (i.e. the bhav~gra). In other words each
dhydna or attainment is subtle in relation to the one below and gross in
relation to the one above -- a process extending to the fourth formless
jhdna. The intended implication is that this is incompatible with the
canonical account where neither vitarka nor vicāra are present in the higher
dhydnas. The Pali tradition avoids this problem by making vitakka and
vicāra gross and subtle in relation to one another rather than the causes of
the mind's grossness and subtlety in general.6~
The Abhidharm~vatdra 61 and the Abhidharmadipa 62 declare that vitarka
is the cause of the activity of the five (sensory) consciousnesses which are
gross, while vicāra is favourable to the activity of mind consciousness
(manovijfidna). They also describe vitarka as differentiating the type of
sensory feature (visayanimittaprakdravikalpin) involved and as having its
activity stirred up by the wind of labelling (samjfid) i.e. it is stimulated by
the constant flow of labelled sensations. Yogficfirin authors give definitions
which are slightly more reminiscent of the Pali Abhidhammapitaka and the
Petakopadesa. It suffices to quote the Abhidharmasamuccaya:
What is vitarka? It is a mental murmuring which searches about (for the object) in
dependence either upon volition (cetand) or upon understanding (praj~6). But that is
grossness of mind. What is vicāra? It is a mental murmuring which investigates (the object)
in dependence either upon volition or upon understanding. But that is subtlety of mind.63148 L.S. COUSINS
STAGES OF SAMADHI IN THE YOGA-SUTRA
The parallelism between the description of sampraj~dta-samādhi in the
Yoga-sutra and the traditional descriptions of the rlipa-jhānas in Buddhist
sources has long been noted.64 Careful examination of the text of the Yogas~tra (Yogas) and its bhdsya suggests that the resemblance is even closer
than has always been appreciated. Two passages are relevant for this
purpose. The first of these (Yogas i 17) reads:
vitarka-vicdr~nanddsmitd-rlip~nugamdt sampraj~dtah. I
It is called [the cessation 65] which clearly comprehends [its
object] as a result of being accompanied by forms 66 of vitarka,
vicāra, bliss and a sense of I.
The commentators are agreed that this indicates four distinct stages,
which may be tabulated, after the bhdsya, as follows:
sense of I
sense of I
sense of I
sense of I
Such a series poses no problems and is, as we shall see, in good agreement
with the Buddhist sources.
The bhdsya to this passage defines vitarka as: cittasy(dambane sth~la
dbhogah. "gross directing of the mind to an object." vicāra is correspondingly subtle. This is quite close to some of the Buddhist definitions we have
seen and strikingly different to the more typical definition of the later
commentaries which sees grossness and subtlety in terms of the object of
the mind rather than as a feature of the mind itself. However this shouldVITAKKA/VITARKA AND VICARA 149
not be overstated. It is certainly a fundamental assumption of both abhidharma and S- .arpkhya-yoga that higher states of consciousness are both
themselves more subtle and possess subtler objects. The point is rather that
in adapting material of ultimately Buddhist origin the Yoga tradition tends
to shift it from the rather psychological bias of abhidharma towards a more
In fact this definition in terms of the subtlety of the object is in part
derived from a passage later in the same chapter of the Yoga-sfara to which
we must now turn:
gabd~rtha-ffidna-vikalpaih, samkirnd sa-vitarkd samdpattih. 
etayaiva sa-vicdrd nirvicdrd ca s~ks.ma-visayd vydkhydtd 
sft~ma-visayatvam c~lihga-paryavasdnam 
42. The attainment with vitarka is associated with concepts arising
from knowledge of the meanings of words.67
43. When mindfulness (smrti) is pure [the attainment[ without
vitarka reveals only the object and is as if empty of its own
44. By this has been explained [the attainment] with vicāra, whereas
[the attainment[ without vicāra may be explained as having subtle
45. Moreover the sphere of subtle objects ends with that which has
no distinguishing marks [i.e. prakrti].
The commentators are a little misleading here. Their discussion is in
terms of the savitarka/nirvitarka, savicāra/nirvicāra terminology, creating a
new set of four samdpattis. If this is taken literally, it creates problems in
reconciling the new set with the version from the bhdsya set out in Table I.
In the light of the Buddhist information that interpretation is almost certainly
correct.69 Yet there is no place for a nirvitarka-samādhi distinct from
According to the commentaries to i 44 savicāra- and nirvicāra-
(samdpatti) are s~ksma-visayd; savitarka- and nirvitarka- are sthala-visayd.
This can be interpreted as a single pair -- these last are gross and all higher
stages of sabifa-samādhi are subtle. However, on further investigation an
alternative solution appears possible. The redactor of the Yoga-sutra may
rather have been thinking in terms of a series of stages. A is gross in150 L. S. COUSINS
relation to the subtlety of B, which is itself gross in relation to the subtlety
of C, yet again gross in relation to D, and so on.
In the light of the general nature of this type of yogic practice, this
second explanation seems much more plausible. Some examples of this kind
of usage are cited in the Visuddhimagga from traditional sources.7~ Here
the first ]hdna is gross (oldrika) where the second is subtle (sukhuma), but
the second is gross where the third is subtle and the third is gross where the
fourth is subtle. Many other examples could be cited from abhidhammic
If we understand the u in this way, we can derive the following
1. sa-vitarka vitarka
sense of I
2. nirvitarka subtle vicāra
gross sense of I
3. nirvicāra bliss
4. *nirdnanda sense of I
This interpretation makes more plausible the translation adopted for Yogas
i 44 (see note 68). Obviously the terms nirvitarka and nirvicāra can also be
taken as referring to all higher levels of samādhi beyond the first (or
second) in the table as these too lack vitarka and vicāra. Equally all higher
levels and mental objects can be described as subtle, not only the one
immediately above. Correspondingly even the lowest level can be described
as possessing vicāra. 71
It should also be mentioned that the influence of the Buddhist account ofVITAKKA/VITARKA AND VICARA 151
the four dhydnas is already evident in one passage of the Mahdbhdrata, as
pointed out by V. M. Bedekar. 72
It is clear that at the time of the formation of many of the classical abhid h a m m a schools in the second and third centuries B.C., several different
lists of factors of samādhi were extant in different suttas. By far the most
important of these was the list contained in the standard formula for the
1. First jhdna -- savitakkam. , savicāram. , vivekajam. , pitisukham. ;
2. Second jhdna -- ajjhattam sampasddanam. ,73 cetaso ekodibhdvam. ,
samādhijam. , pitisukham. ;
3. Third jhdna -- upekkhako viharati, sato ca sampajdno sukha~ ca
kdyena patisam, vedeti, upekkhako, satired, sukhavihdrf ;
4. Fourth jhdna - adukkham. , asukham. , upekkhdsatipdrisuddhi.
The Vibhahga, an early canonical abhidhamma work, formulates (in its
Suttantabhdjaniya) the factors which make up each jhdna as follows: TM
joy and happiness
one-pointedness of mind75
joy and happiness
one-pointedness of mind
mindfulness (sati )
one-pointedness of mind
one-pointedness of mind
The list given in the Abhidharmakoga differs only for the fourth dhydna,
where it has: 1. neutral feeling; 2. upeks..dparisuddhi; 3. smrtiparisuddhi; 4.
samādhi. TM152 L.S. COUSINS
Also in early sources was a division of samādhi into three k i n d s "77
1. With vitakka and vicāra;
2. Without vitakka but with vicāra;
3. With neither vitakka nor vicāra.
The list posed problems for systematic analysis, as the second kind cannot
be precisely equated with any one of the four jhānas. Early schools
resolved the probem in one of two different ways. The Sarvfistivfidins and
others introduced an intermediate stage or dhydn~ntara between the first
and second dhydnas. The Vibhajjav~dins/Theravfidins preferred to adopt an
alternative list of five jhānas for purposes of abhidhamma analysis and
distributed the factors as follows:78
one-pointedness of mind
one-pointedness of mind
one-pointedness of mind
one-pointedness of mind
one-pointedness of mind
Another ancient list divided similar material into three sections: 79
samādhi with vitakka and vicāra, with only vicāra, with neither;
samādhi with and without joy;
samādhi with s&a or with equipoise.
It seems clear that two distinct stages are to be found in the sources. In the
early material we find a less standardized situation with the experience ofVITAKKA/VITARKA AND VICARA 153
samādhi at the centre of the stage and the classification into types more
varied and perhaps more experiential in nature. Only later arises the
attempt to harmonize this material into a single list, as is done in the
Sarvfistiv~tdin tradition. In this respect the Therav~din device of an alternate
set of five jhānas preserves more faithfully an earlier variety,s~
For the canonical abhidhamma, vitakka at its weakest results in a tendency
to speculate and fix upon ideas. More strongly developed it is the ability to
apply the mind to something and to fix it upon a (meditative) object. vicāra
at its weakest is simply the tendency of the mind to wander. More highly
developed it is the ability to explore and examine an object. In one way we
can say that vitakka is 'thinking of' something, whereas vicāra is 'thinking
about' that same thing, but in fact the latter is probably intended to refer
more to what we would now describe as the mind's associative faculty.81
A samādhi with vicāra but without vitakka is a state of mind in which
the mind freely associates about a meditative object without deliberately
thinking of anything specific. It must obviously be preceded by one with
vitakka. If it were not, then the mind would simply wander to any kind of
object and soon lose (or never acquire) the kind of focus and unification
required for samādhi. The situation is different once a samādhi with
vitakka is sufficiently strongly developed. Then vitakka can drop away
because that focus is well enough established not to need further reinforcement. 82
The later texts do not depart radically from this understanding. Such
differences as we do find are simply due to the need to conform to the
requirements of theory. For the Vaibh~tsikas and Buddhaghosa this means
conformity to strict momentariness and precise definition as distinct
dharmas. For the Yoga-sfitra, and even more for its commentaries, the
stages of samādhi need to be related to the levels of S ~ . khya evolution.
Abbreviations of names of texts are those used in the Critical Pali Dictionary unless
1 A useful presentation of part of the case for such influences in: Johannes Bronkhorst, The
Two Traditions of Meditation in Ancient India, Stuttgart, 1986, chapter six.
2 In Philip Denwood and Alexander Piatigorsky, Buddhist Studies -- Ancient and Modern,
London, 1983, pp. 1--11. Some of the material in the first part of this article was included
in a draft version of that paper.154 L. S. C O U S I N S
3 M III 73.
4 D I 16; 21; 23; 29; a similar formula at M I 68, etc.; 520.
5 A I 89, etc.; 193 ft.; 11 191 f.
6 D I 12 etc.
7 M I 487; 1I 172 f.; A II 189 f.; Sn 885 f.
8 D I I 3 6 f . ; M I 1 6 7 ; S I 1 3 6 ; V i n I 4 .
9 Sv-pt I 188 ft.; Sv I 106 f.
10 E.g. Dhs-a 114.
11 M I 114 ft.; M I I I 114; 125; S I 203; II 151 ft.; 1/I 93; A I 148; 254; 275; II 16; 76;
117; 252; HI 390; 428 f.; 446; cf. also M I 118; 133; III 129 ft.; and gehasita-vitakka at M I
124; S I 186.
12 D II 277; M I I I 124; S II 151 ft.; 153 f.; A IV 147; safifid and vitakka are juxtaposed at
M I 133; A III 428; cf. also M I I I 129.
a3 The eight thoughts of a great man D HI 287, A IV 229; not thinking a thought connected
with the body while dwelling in contemplation of the body M III 136; thinking whatever
thought one desires to think, etc. A II 36 f.; the monk who is full of thought and spends his
day in thoughts about dhamma A III 87; mindfulness of in and out breathing should be
developed in order to cut off thought A IV 353; subtle defilements are thoughts connected
with clan, country or reputation -- after these go, dhamma thoughts remain A I 254;
thought is the world's vicāran~ S I 39 f. Some of these could be interpreted in more
technical senses. Evil unskilful thoughts at A I 280 probably refers to the three kinds of
14 Thought reading D I I I 104, A I 171, cf. D I 213; where vitakka and vicāra cease and
who dwells constantly making them cease A IV 411; a place is oppressive in which they
have not ceased A IV 550; sah~dmanasik~ra associated with vitakka are an illness for
someone dwelling in the second jhdna A IV 415; the jhdyin without vitakka S I 186;
vitakka and vicflra have ceased, are tranquilized and calmed in the second jh~na S IV 217,
cf. A IV 409; they are movement in (the peace of) the first jhdna M 1 454, cf. D I 37; also
in lists of types of samādhi, see notes 77 and 79 below.
15 M I 30I; S IV 293.
16 A IV 385.
17 A II 36 f.
18 D III 215; M II 24 ft.; A V 31; kdma-Saṇkappa alone at A 1]I 259, 145--6. The three
unskilful sa~kappas are probably intended at Sn 818 and A I 281 -- in the last case
satikappa in verse corresponds to vitakka in prose.
19 Due to lack of mindfulness M I 453 f., S IV 190; due to not guarding the senses S IV
76 f., 136 f.; similar at M III 132 and 136 (gehasita).
20 D III 42--6; M I 192 if.; IlI 276 f. (here the meaning 'purpose' is required); A V 92; 94,
97; 99; 104.
21 S II 143 f.; 144 ft.; 146 f.; 147 ft.; cf. also M II 27 f.
22 Pet 168.
23 A IV 199; 202; also Vin II 238 f.; Ud 53; 55; S II 184; Ap 23.
24 Vin II 136 f.; I[I 217; 257 f.
25 Vibh 195; 197; 202; 254; 259. Probably appami as the name for the concluding phrases
in Dhs should be understood in the same way.
26 Dhs-a 142; Ps IV 132 f.
27 Dhs-a 142 f.
28 Nett 33; Pet 92.
29 Paris I 16; 17; etc.
30 Patis I 79.
31 M III 251; D II 312; S V 9; etc.V I T A K K A / V I T A R K A A N D V I C . ~ R A 1 5 5
32 M I 402.
33 S IV 189 f.; Paris II 19.
34 A IV 147; cf. Vibh 103 f.
35 E.g. M I 115 f.: there is an inclination of the mind towards whatever one frequently
applies vitakka and vicāra to; cf. also M I 116: if I apply vitakka and vic6ra for too long,
my body would become tired; M I 1 4 4 : . . . applies vitakka and vicF~raby night to the affairs
of the day; M II 253 f . : . . , applies vitakka and vicāra in accordance with that teaching.
36 A III 87 f.; 177 f.; 361 f.; 382 f.; IV 86; A III 21 ft. = D III 242 (i.e. the Sa~gftipariy6ya);
the exception is A I 264.
37 Cf. BHSD s.v. upavicāra and A IIl 363 f.
38 D IH 244 f.; M IN 216 f.; S IV 232; A 1 176; cf. Vibh 381.
39 Dhs-a 143.
40 A IV 47--51; E e and C e 1970 read anusandati; the latter cites a reading anusanthdti cf.
Skt anusamsthti 'to follow'.
4~ PTC lists twenty seven occurrences s.v. dhata-.
4l See for example: It 72; Nidd I 386; 493; 501 Patis I 17; 36; 136; 178 f.; Vibh 86 f.;
103 f.; 362; Kvu 412 ff. For takka: It 37; Nidd I 293 f.; 360; 400; 482; 498; 501; II 29;
43 Pet 142.
44 /q~namoli (Pet Trsl. 190) so emends apeti.
45 Sv I 144; Dhs-a 115; Vista 142.
46 This paragraph must be compared to Nett 19--20 where there are a number of parallels.
Probably the comparison in Pet of vitakka with aparihhd and of vicāra with parih~d should
read abhih~d and parih~d (or vice versa).
47 So correct/q~mamoli's translation.
4s Mil 62 f.
49 E e reads anusandahati. S e (cited CPD) reads anusandati. Dhs-a has anusadddyati which I
follow in the translation. There may be, or have been thought to be, some connection with
the anusandhanat~ of the Dhammasa~gani register but it is difficult to make sense of this.
Possibly both were taken to mean 'continuing in sequence' cf. BHSD s.v. anusamdhi.
50 Vim Trsl. 86 ff.
51 Nett Trsl. liv.
52 Sp I 144; Dhs-a 114 f.; Vism 142 f. See also the Chinese version of Sp (trsl. P. V. Bapat
and A. Hirakawa, 1970) p. 104 and Patis-a 1 181 f.; Nidd-a I 127 f.; Moh 12.
53 For the bell simile, see also N. Aiyaswami Sastri, Satyasiddhig~stra of Harivarman,
Baroda, 1975 and 1978, I 165 and 216; II pp. 134 and 186.
54 Sv I 63; Ps I 48; Mp I 106; cf. Sv I 121 f.; Ps II 348; Mil 62 and note the earlier pair at
Patis I 17.
55 "The Visuddhimaggga (also Patis-a I 182 and Nidd-a I 128) cites instead the
Dukanip~itatthakath6, but this too must be the old Sinhala commentary as it is not in Mp.
56 Abhidh-k-bh II 33.
57 Marcel Van Velthem, Le Trait~ de la Descente darts la Profonde Loi de l'Arhat
Skandhila, Louvain, 1977, p. 16. Further material cited in SWTF s.v. auddrika-paticavijhdna-hetu-dharman (kindly sent to me by Dr. Siglinde Dietz).
5s ~tnti Bhiksu ~stri, Jtiimaprasthana-~dstra of Kdtydyanfputra retranslated into Sanskrit
from the Chinese version of Hiuan Tsang, g~ntiniketan, 1955, p. 53.
59 Charles Willemen, The Essence of Metaphysics, Brussels, 1975, p. 27. Also p. 106 where
vitakka is explained as "when, at the moment of engaging in concentration, one begins to
produce wholesome qualities, one first reflects with coarse thoughts," while for vicāra we
have: '%vhen one connects and links the thoughts with subtlety."
60 It must be noted, however, that the Vaibh~sikas probably did not intend to go so far.
This is simply the implication claimed by Vasubandhu in their use of the simile of ghee in156 L . S . C O U S I N S
water to defend the simultaneous presence of both dharmas. Note that Vasubandhu, but not
the Sautrgntikas, denies the possibility of such a simultaneous occurrence since for him they
are two degrees of a single dharma.
61 Van Velthem, op. cit., p. 16.
62 p. S. Jaini, Abhidharmadfpa with Vibhdsaprabhdvrtti, Second Edition, Patna, 1977, p. 81.
Jaini's footnotes give a number of relevant source passages. See also the general discussion
of vitarka and vicāra in the Introduction pp. 83--88. Further comments and references in:
Stefan Anacker, Seven Works of Vasubandhu, Delhi, 1984, p. 77 f.
63 Cited Jaini, op. cit., p. 81 f. Some similar passages also in the SatyasiddhigOstra --
Aiyaswami Sastri, op. cit., Vol II p. 185 f. cf. pp. 134 f.; 173; 246 f.; 367 ft.; 385 ff.
64 Bronkhorst, op. cir.
65 Supplying the noun nirodhah from Yogas i 2 and i 12. Alternatively samprajhdtah. (sc.
samādhih) cf. Yogas-bh and Bhoja.
66 Just possibly nipa here refers to the object of samādhi, as in some Buddhist sources. See
L. S. Cousins, "Buddhist jhdna. Its nature and attainment according to Pali sources,"
Religion III (1973) p. 119. We would then translate: "... as a result of following a [mental[
object with vitarka ..."
67 Following Shlomo Pines and Tuvia Gelblum, "Al-birfmi's Arabic Version of Patafijali's
Yogasfitra: A translation of his first chapter and a comparison with related Sanskrit texts,"
BSOAS XXIX (1966) p. 325 n. 234. This is a translation of the text in isolation. If one
prefers to take account of the use of these terms elsewhere in Yogas (i 9 and iii 17), then
the following would be preferable:
42. The attainment with vitarka is mixed with ]erroneous[ identification of the
sound of the word with the object [to which it refers] and the general concept
[implied by the word].
68 Following the alternative translation given in Pines and Gelblum, op. cit., p. 325 n. 234.
See Table II. The more usual translation is:
44. [The attainments[ with and without vicāra, which have subtle objects, are
explained in just the same way.
69 The four factors given by the Yoga-stitra are simply a modification of the various
Buddhist fists of factors. Ananda is in any case a synonym for prfti and asmitd is substituted
for upeksfi/tatramaffhattatd with an eye to S- .mpkhya theory (buddhi) -- even at the price of
some inconsistency with Yogas ii 3.
70 Vism 275.
71 As with Yogas i 17 above.
72 "The Dhy~mayoga in the Mah~bh~trata (XII 188)," Bhdrat~ya Vidyd, 20--21 (1960--61)
116--25. Further discussed in Bronkhorst, op. cit., p. 65 ft.
73 Compare the adhydtma-prasdda of Yogas i 47.
74 Vibh 257, line 30; etc.
75 The addition of cittekaggatd ~ samdMhi to the list is explicit in such contexts as
M I 294 f.; III 25.
76 Abhidh-k-bh VIII 8; the Abhidharmasamuccaya (p. 68) is similar; cf. Pet 139;
M I l I 25 L; cp. also D I 196.
77 DN UI 219 = 274 # SN IV 360 # 362 f.
78 Dhs 33 ft.
79 M m 162 = A IV 300.
80 A different view: Walpola Rahula, "A Comparative Study of Dhy~mas according to
Theravada, Sarv~stivfida and MahhySna," The Maha Bodhi, June 1962, pp. 190--99.V I T A K K A / V I T A R K A A N D V I C . A R A 157
81 It is not as such 'sustained application of mind' nor is it 'holding the mind' on an object
- - these are results of vicāra, not its nature.
82 The Dhs-mt (Ce 1938) 166 (to Dhs-a 115) describes vitakka in access or absorption
samādhi as unwavering (niccalo hutvd) and as having entered into the object (anupavisitvd
pavattati). The Anutikd explains that when it is active continuously on a single object it does
not vibrate as it does with a new object.
University of Manchester