THE KATHA-UPANISHAD is probably more widely known than
any other Upanishad. It formed part of the Persian translation,
was rendered into English by Râmmohun Roy, and
has since been frequently quoted by English, French,
and German writers as one of the most perfect specimens
of the mystic philosophy and poetry of the ancient Hindus.
It was in the year 1845 that I first copied at Berlin
the text of this Upanishad, the commentary of Sankara
(MS. 127 Chambers), and the gloss of Gopâlayogin
(MS. 224 Chambers). The text and commentary of Sankara
and the gloss of Ânandagiri have since been edited
by Dr. Roer in the Bibliotheca Indica, with translation
and notes. There are other translations, more or less
perfect, by Râmmohun Roy, Windischmann, Poley,
Weber, Muir, Regnaud, Gough, and others. But there still
remained many difficult and obscure portions, and I
hope that in some at least of the passages where I differ
from my predecessors, not excepting Sankara, I may have
succeeded in rendering the original meaning of the author
more intelligible than it has hitherto been.
The text of the Katha-upanishad is in some MSS. ascribed
to the Yagur-veda. In the Chambers MS. of the commentary
also it is said to belong to that Veda , and in the
Muktikopanisbad it stands first among the Upanishads
of the Black Yagur-veda. According to Colebrooke (Miscellaneous
Essays, 1, 96, note) it is referred to the Sâma-veda
also. Generally, however, it is counted as one of the
The reason why it is ascribed to the Yagur-veda, is
probably because the legend of Nakiketas occurs in the
Brâhmana of the Taittirîya Yagur-veda. Here
we read (III, 1, 8):
Vâgasravasa, wishing for rewards, sacrificed
[1. MS. 133 is a mere copy of MS. 127.
2 Yagurvede Kathavallîbhâshyam.]
wealth. He had a son, called Nakiketas. While he was
still a boy, faith entered into him at the time when
the cows that were to be given (by his father) as presents
to the priests, were brought in. He said: 'Father, to
whom wilt thou give me?' He said so a second and third
time. The father turned round and said to him: 'To Death,
I give thee.'
Then a voice said to the young Gautama, as he stood
up: 'He (thy father) said, Go away to the house of Death,
I give thee to Death.' Go therefore to Death when he
is not at home, and dwell in his house for three nights
without eating. If he should ask thee, 'Boy, how many
nights hast thou been here?' say, 'Three.' When he asks
thee, 'What didst thou eat the first night?' say, 'Thy
offspring.' 'What didst thou eat the second night?'
say, 'Thy cattle.' 'What didst thou eat the third night?'
say, 'Thy good works.'
He went to Death, while he was away from home, and
lie dwelt in his house for three nights without eating.
When Death returned, he asked: 'Boy, how many nights
hast thou been here?' He answered: I Three.' 'What didst
thou eat the first night?' 'Thy offspring.', 'What didst
thou eat the second night?' 'Thy cattle.' 'What didst
thou eat the third night?' 'Thy good works.'
Then he said: 'My respect to thee, O venerable sir!
Choose a boon.'
'May I return living to my father,' he said.
'Choose a second boon.'
'Tell me how my good works may never perish.'
Then he explained to him this Nâkiketa fire (sacrifice),
and hence his good works do not perish.
'Choose a third boon.'
'Tell me the conquest of death again.'
Then he explained to him this (chief) Nâkiketa
fire (sacrifice), and hence he conquered death again
This story, which in the Brâhmana is told in
order to explain the name of a certain sacrificial ceremony
[1. The commentator explains punar-mrityu
as the death that follows after the present inevitable
Nâkiketa, was used as a peg on which to hang
the doctrines of the Upanishad. In its original form
it mayhave constituted one Adhyâya only, and the
very fact of its division into two Adhyâyas may
show that the compilers of the Upanishad were still
aware of its gradual origin. We have no means, however,
of determining its original form, nor should we even
be justified in maintaining that the first Adhyâya
ever existed by itself, and that the second was added
at a much later time. Whatever its component elements
may have been before it was an Upanishad, when it was
an Upanishad it consisted of six Vallîs, neither
more nor less.
The name of vallî, lit. creeper, as a subdivision
of a Vedic work, is important. It occurs again in the
Taittirîya Upanishads. Professor Weber thinks
that vallî, creeper, in the sense of chapter,
is based on a modern metaphor, and was primarily intended
for a creeper, attached to the sikhâs or branches
of the Veda. More likely, however, it was used in
the same sense as parvan, a joint, a shoot, a branch,
i.e. a division.
Various attempts have been made to distinguish the
more modern from the more ancient portions of our Upanishad.
No doubt there are peculiarities of metre, grammar,
language, and thought which indicate the more primitive
or the more modern character of certain verses. There
are repetitions which offend us, and there are several
passages which are clearly taken over from other Upanishads,
where they seem to have had their original place. Thirty-five
years ago, when I first worked at this Upanishad, I
saw no difficulty in re-establishing what I thought
the original text of the Upanishad must have been. I
now feel that we know so little of the time and the
circumstances when these half-prose and half-metrical
Upanishads were first put together, that I should hesitate
[1. History of Indian Literature, p.
93, note; p. 157.
2. Though it would be unfair to hold
Professor Weber responsible for his remarks on this
and other questions connected with the Upanishads published
many years ago (Indische Studien, 1853, p. 197), and
though I have hardly ever thought it necessary to criticise
them, some of his remarks are not without their value
before expunging even the most modern-sounding lines
from the original context of these Vedântic essays.
The mention of Dhâtri, creator, for instance
(Kath. Up. II, 20), is certainly startling, and seems
to have given rise to a very early conjectural emendation.
But dhâtri and vidhâtri occur in the hymns
of the Rig-veda (X, 82, 2), and in the Upanishads (Maitr.
Up. VI, 8); and Dhâtri, as almost a personal deity,
is invoked with Pragâpati in Rig-veda X, 184,
I. Deva, in the sense of God (Kath. Up. II, 12), is
equally strange, but occurs in other Upanishads also
(Maitr. Up. VI, 23; Svetâsv. Up. I, 3). Much might
be said about setu, bridge (Kath. Up. III, 2; Mund.
Up. II, 2, 5), âdarsa, mirror (Kath. Up.VI, 5),
as being characteristic of a later age. But setu is
not a bridge, in our sense of the word, but rather a
wall, a bank, a barrier, and occurs frequently in other
Upanishads (Maitr. Up. VII. 7; Khând. Up. VIII,
4; Brih. Up. IV, 4, 22, &c.), while âdarsas,
or mirrors, are mentioned in the Brihadâranyaka
and the Srauta-sûtras. Till we know something
more about the date of the first and the last composition
or compilation of the Upanishads, how are we to tell
what subjects and what ideas the first author or the
last collector was familiar with? To attempt the impossible
may seem courageous, but it is hardly scholarlike.
With regard to faulty or irregular readings, we can
never know whether they are due to the original composers,
the compilers, the repeaters, or lastly the writers
of the Upanishads. It is easy to say that adresya (Mund.
Up. I, 1, 6) ought to be adrisya; but who would venture
to correct that form? Whenever that verse is quoted,
it is quoted with adresya, not adrisya. The commentators
themselves tell us sometimes that certain forms are
either Vedic or due to carelessness (pramâdapâtha);
but that very fact shows that such a form, for instance,
as samîyâta (Khând. Up. I, 12, 3)
rests on an old authority.
No doubt, if we have the original text of an author,
and can prove that his text was corrupted by later compilers
[1. See Regnaud, Le Pessimisme Brahmanique,
Annales du Musée Guimet, 1880; tom. i, p. 101.]
or copyists or printers, we have a right to remove
those later alterations, whether they be improvements
or corruptions. But where, as in our case, we can never
hope to gain access to original documents, and where
we can only hope, by pointing out what is clearly more
modem than the rest or, it may be, faulty, to gain an
approximate conception of what the original composer
may have had in his mind, before handing his composition
over to the safe keeping of oral tradition, it is almost
a duty to discourage, as much as lies in our power,
the work of reconstructing an old text by so-called
conjectural emendations or critical omissions.
I have little doubt, for instance, that the three verses
16-18 in the first Vallî of the Katha-upanishad
are later additions, but I should not therefore venture
to remove them. Death had granted three boons to Nakiketas,
and no more. In a later portion, however, of the Upanishad
(II, 3), the expression srinkâ vittamayî
occurs, which I have translated by 'the road which leads
to wealth.' As it is said that Nakiketas did not choose
that srinkâ, some reader must have supposed that
a srinkâ was offered him by Death. Srinkâ,
however, meant commonly a string or necklace, and hence
arose the idea that Death must have offered a necklace
as an additional gift to Nakiketas. Besides this, there
was another honour done to Nakiketas by Mrityu, namely,
his allowing the sacrifice which he had taught him,
to be called by his name. This also, it was supposed,
ought to have been distinctly mentioned before, and
hence the insertion of the three verses 16-18. They
are clumsily put in, for after punar evâha, 'he
said again,' verse 16 ought not to have commenced by
tam abravît, 'he said to him.' They contain nothing
new, for the fact that the sacrifice is to be called
after Nakiketas was sufficiently indicated by verse
19, 'This, O Nakiketas, is thy fire which leads to heaven,
which thou hast chosen as thy second boon.' But so anxious
was the interpolator to impress upon his hearers the
fact that the sacrifice should in future go by that
name, that, in spite of the metre, he inserted tavaiva,
'of thee alone,' in verse 19.