मानसार शिल्पशास्त्र (प्रसन्ना कुमार आचार्य - प्राचीन भारतीय स्थापत्य, नगर-नियोजन व मूर्ति-शिल्प महाकोश)

MANASARA
ON
ARCHITECTURE AND SCULPTURE
SANSKRIT TEXT WITH CRITICAL NOTES
EDITED BY
PRASANNA KUMAR ACHARYA, I.E.S.
കേരള സാഹിത്യ
അക്കാദമി

കേരള സാഹിത്യ
അക്കാദമി
RAMA VARMA RESEARCH INSTITUTE,
TRICHUR, COCHIN STATE.

RAMA VARMA RESEARCH INSTITUTE,
TRICHUR, COCHIN STATE.
MĀNASĀRA
ON
ARCHITECTURE AND SCULPTURE

Five Volumes with an Encyclopædia, Introduction, Text, Translation,
and 157 Plates in line and in colours
ON
HINDU ARCHITECTURE
Volume I—A Dictionary of Hindu Architecture
Volume II—Indian Architecture according to Mānasāra Śilpaśāstra
Volume III—Mānasāra—Sanskrit Text with Critical Notes
Volume IV—Architecture of Mānasāra—Translation in English
Volume V—Architecture of Mānasāra—Plates I—CXXXV (Architectural)
CXXXVI—CLVII (Sculptural)

MĀNASĀRA
ON
ARCHITECTURE AND SCULPTURE
SANSKRIT TEXT WITH CRITICAL NOTES
EDITED BY
PRASANNA KUMAR ACHARYA, I.E.S.
M.A. (CALCUTTA), PH.D. (LEYDEN), D.LIT. (LONDON)
Dean of the Faculty of Arts, Head
of the Oriental Departments, Professor
of Sanskrit, University of Allahabad
BLISHED BY THE OXFORD UNIVERSITY PRESS
LONDON NEW YORK BOMBAY
CALCUTTA MADRAS
CHINA JAPAN

കേരള സാഹിത്യ
അക്കാദമി
PRINTED BY THE SUPERINTENDENT, PRINTING AND STATIONERY, UNITED PROVINCES, ALLAHABAD, INDIA

"What the learned world demand
of us in India is to be quite
certain of our data, to place the
monumental record before them
exactly as it now exists, and to in-
terpret it faithfully and literally."

PREFACE
History of Publication
IN 1834, in his Essay on the Architecture of the Hindus, Rām Rāz
referred to the contents of the first few chapters of the Mānasāra
from a fragmentary manuscript he had access to. Since then more
manuscripts have been discovered, catalogued, and brought to the
notice of the public. But the Sanskrit scholars like Dr. G. Bühler and
Sir R. G. Bhandarkar have authentically branded this class of literature
as written in “most barbarous Sanskrit.”¹ Besides, the technical
terms and obsolete expressions of the Śilpa-śāstra, practically a lost
science, have had no place in any of the existing dictionaries. The
texts contain no drawings, diagrams, or sketches. The then buildings,
religious, military, especially residential, have mostly been destroyed,
partly by time and partly by foreign invaders. Obviously these diffi-
culties stood in the way of the elderly scholars and nobody had made
any attempt to deal with this huge text for over 80 years, when, being
induced by youthful enthusiasm, the present writer undertook in 1914
the task of giving “rebirth” to this “monumental work.”
The unsatisfactory condition, both original and accidental, of
the existing manuscripts of the śilpa-śāstra generally, and of the
Mānasāra in particular, aggravated the expedition into this unexplored
field of Sanskrit learning. The unknown ways have had to be
discovered at great pains and many new things had to be learnt after
bitter experience and elaborate experiments. The vast store of
literature on the subject and the inscriptions and other archaeological
records have had to be ransacked. Extensive tours in company with
trained architects and engineers have had to be undertaken, involving
great expenditure of time, money, and convenience from Landikotal
on the borders of Afghanistan to Rameswaram in the south, from
¹ Ep. Ind., Vol. 1, page 377 ; Ind. Ant., Vol. XII, page 140.

x PREFACE
Mahenjadaro in Sindh to Shillong in Assam, in order to examine ancient
monuments throughout the country, including the cave temples of
Western and Central India, the monuments of the Indian States in
Rajputana and in Hyderabad, the hill temples of Kedarnath,
Badrinath, Tunganath, etc. An encyclopædic dictionary has had to be
compiled wherein for the first time an attempt has been made to settle
the meanings of, and explain in English, some three thousand technical
terms by illustrative quotations from cognate literature. In addition,
some two hundred measured drawings, diagrams, and sketches in line
and in colours have had to be prepared at great trouble and expense.
Thus the peculiar pain of giving “ rebirth ” to a once fully grown
“ barbarous ” child of unknown origin has been brought home to the
writer. The youthful enthusiasm has been put to a great trial through
the continuous struggle for seventeen years. With great diffidence,
therefore, this curious child is presented to the learned world. If
received with sympathy and kindness even such a child may thrive
and bring some solace to everybody responsible for and interested in
it. Ultimately it may confer a lasting benefit on the country and the
nation, because our architectural policy of the past few hundred years,
based as it has been on foreign imitation, and in an entirely different
climate and soil, has not proved quite successful in regard to temples
and humble dwelling-houses, if not in regard to public edifices also.
Lord Pentland, the then Governor of Madras and in charge of the
Education Department, was induced by Sir John Marshall, lately the
Director-General of Archaeology in India, to take an interest in the
Mānasāra, and some preliminary arrangements were made for its publi-
cation from Madras. But the unfortunate coincidence of His
Excellency's retirement and Sir John's absence from India at the
time of my arrival in Madras, forced me to leave the Presidency for
Simla, where all the endeavours and the sympathy of Sir (then Mr.)
George Anderson, Dr. D. B. Spooner, Sir Henry Sharp and Sir C.
Sankaran Nair failed to secure a situation suitable for the “ rebirth ” of
the Mānasāra. At last an opportunity arose when the work could be
proceeded with in spite of other exacting duties. The Government of

PREFACE xi
India recommended, and the Government of the United Provinces
offered me a post in the Indian Educational Service at Muir Central
College, Allahabad. Sir Claude F. de la Fosse, the then Director of
Public Instruction and the first Vice-Chancellor of the reorganized
Allahabad University, took up the matter with friendly sympathy
and scholarly interest and induced the great educationist Governor,
Sir Harcourt Butler, to sanction the publication on behalf of the
Government of the United Provinces of Agra and Oudh through the
Oxford University Press.
Description of the Manuscripts
The manuscripts which have been used in preparing the text of
this First Edition of the Mānasāra are the following :—
(A) This belongs to the Library of the India Office (B.482).
It is written in Devanāgarī script on modern paper
9¾" by 7½". It contains 100 folios, each page of which
has invariably 21 lines, except the last page which has
only 1 line. Each line is half a śloka. The ślokas are
inaccurately numbered from the beginning to the end.
The writing seems to be by one and the same hand.
The name of the writer is not known. The colophons
are written in red ink. The manuscript gives up the
Mānasāra at folio 78b. line 16, the middle of Chapter
XVIII (Vimāna-lakshaṇa), without any colophon. The
remaining 22 folios, apparently belonging to another
work, are retained in Appendix क. This portion has
two colophons—folio 93a, iti Sūkshma-śāstre Kāma-
koshṭa-paṭalaḥ ; folio 95b, iti Padma-saṁhitāyāṁ
Raurava-bhede Kāma-koshṭa-lakshaṇa-vidhiḥ paṭalaḥ. On
page 101 it breaks off abruptly at the middle of line 1.
(B) This has been noticed by Aufrecht in the Catalogues Cata-
logorum, part III, page 97. It belongs to the Oriental
Library of the Asiatic Society of Bengal. It is written
on paper 12" by 8½". The first part, to folio 88, is a

xii PREFACE

different work called Nara-pati-vijaya. The second part,
called Mānasāra-vāstu and also Kumāra-vāstu, contains
163 folios, each page of which has 16 to 24 lines. Each
line contains about three-quarters of a śloka. The ślokas
are not numbered; but on the last page the total is given
as 5,760 and this figure is not, of course, verified by actual
counting. This has been undoubtedly written by more
than one hand. But it is recorded on the last page that
this work has been written by one Rāmanuj chārya for the
Kumpāni bāhādur (the East India Company) in the
Śālivāhana Śaka era 1677. The date of its being recorded
is given in English, 14th April, 1823. The last page
contains the mark of a seal in three different scripts—
Urdu, Nāgarī, and Bengali. It begins with Chapter XV
(Pāda or Stambha), and folio 108b has the colophon of
the last chapter. The remaining folios 108 to 163 contain
the first fourteen chapters.
(C) This belongs to the Library of the India Office (E. 3141,
3014) and has been noticed by C. P. Brown in Eggeling's
Catalogue, page 1131. Dr. Goldstücker refers to it
in his Dictionary, page 284. It is written in Devanāgarī
character on paper 11" by 7". It contains 472 folios,
one page of which is left blank. Each page has 16 to 18
lines, each of which is a half śloka. Ślokas are numbered
up to folio 100. The colophons are not numbered in the
latter part. Most of the chapters are incomplete and des-
cribed in two places. The first sixteen chapters are
altogether lost. The first few pages are quite frag-
mentary. The fly-leaf contains the following: "This
portion of the शिल्पशास्त्र was copied from a very ancient
manuscript in the old Tamil character, discovered at
Trichinopoly. It was imperfect both at the beginning
and at the end. It was written out under the direction
of Charles Philip Brown, 1830." But "from various

PREFACE xiii
indications it would, however, rather seem that the
volume was copied from two different manuscripts, one a
Grantha and the other a Telegu manuscript (Grantha
manuscript, no. 1987, and Telegu manuscript no. 525 in
Taylor's Catalogue Rais., I 71; II 355)." It seems to
have been copied from different sources by different
writers, but evidently not from the sources mentioned
here.
(D) This belongs to the Library of the India Office (Mackenzie
Collection, no. III, 187). It is written in Telegu
character on palm leaf, 13" by 2". It contains 162
folios, each page of which has 9 to 11 lines. Exactly
like 'B,' it begins with Chapter XV and after the
colophon of the last chapter of the book takes up the
first fourteen chapters. The name of the copyist is not
known. The date is given as Śaka 1656 (A.D. 1734).
The characters are very small and indistinct. After
careful collation with 'B,' it seems that both 'B' and
'D' are copies of the same original. 'B' may have
been already copied from 'D' when the latter was
removed to the India Office.
(E) This belongs to the Library of the India Office (Mackenzie
Collection, no. III, 190). It is written in Telegu character
on palm leaf, 19" by 1¾". It contains 68 folios, each
page of which has 7 to 8 lines. But there are some
pages quite blank. Each verse is followed by its
Telegu translation. The pages are numbered 131-198.
It begins at page 131. The colophon reads iti Maya-
mate Vāstu-śāstre Vimāna-lakshaṇa (158, 6, 6). As
Dr. Rost remarks, "the Sanskrit text is hopelessly
corrupt."
(F) This belongs to the Tanjore Palace Estate. It is referred
to in Aufrecht, part I, page 452, part III, page 97,
and in Burnell, page 62, no. 11081. It is written in

Grantha and Tamil characters on Cadjan (palm) leaf,
16½" by 5". The breadth of a single leaf is one inch.
It contains 202 leaves, each of which has 4 to 5 lines.
It can conveniently be divided into three parts. The
first part, in Grantha character, is exactly the same as
'A,' showing their common origin. They are the same
line for line; both break off the Mānasāra abruptly
at the same line. The second part, in Tamil, called
Kāma-koshṭha-bandha, corresponds to Appendix क and
seems to be a commentary on the latter portion of
the first part, which does not belong to the Mānasāra.
The third part contains some chapters belonging to a
different work, probably the Mayamata. Each verse is
accompanied by a Tamil commentary throughout. This
part is noticed in Appendix ख. 'F' like 'A' has,
therefore, only the first seventeen chapters of the Māna-
sāra and a portion of Chapter XVIII.
(G) This belongs to the Tanjore Palace Estate. It is referred
to in Aufrecht, part I, page 452, part III, page 97, and in
Burnell, page 62, no. 11080. It is written in Grantha
character on Cadjan (palm) leaf, 16" by 2¼". The
breadth of a single leaf is 1½". It contains 82 leaves,
each page of which has 4 to 6 lines. The first two
ślokas are invocations to the gods Gaṇeśa and Śiva.
Then come 9 lines from Chapter XXXIV (Maṇḍapa).
The manuscript begins practically with the next Chapter
XXXV (Śālā) and runs to Chapter L (Bhūshaṇa).
(H) This belongs to the Deccan College, Poona, Bombay. It
is referred to in Aufrecht, part III, page 97, and in
Dr. Bhandarkar's Report for the years 1887-91. It is
written in Devanāgarī character on foreign bluish paper,
with water lines, 9½" by 7⅜". It contains 44 leaves,
each page of which has 21 lines. Each line has about
21 letters. The characters are legible, clear and bold, and

PREFACE xv
the writing is fairly correct. The manuscript is in
the form of a modern book, with a margin of one inch
to three-fourths of an inch with a double line of border
in red ink on all sides. The same ink is used to denote
the beginning and the end of the different Chapters, as
also to show the beginning and the conclusion of the
anuṣṭubhs, their pādas, and of sentences. No date
is given. The author is not known. It begins with a
table of contents. Like 'G,' it has only Chapters
XXXV to L. The manuscript itself seems to be a
modern copy of an older manuscript. The page '1a'
has the words "Mānava-Śilpa" as the title of the work,
though in Bhandarkar's report the name is given as
"Mānava-sāra-vāstu-śāstra."
(I) This belongs to the Government Oriental Library, Madras.
It is referred to in Aufrecht, part II, page 104. It is
written in cursive Grantha character on European
paper and bound in size 13¼" by 8". It consists of 495
pages, each of which contains 18 lines. It retains all
the chapters of the Mānasāra in their proper order with
a few exceptions. The copyist gives his name as Loka-
nātha Śāstrī. The date of transcription is not given.
(J) This belongs to the Government of Madras. It is referred
to in Aufrecht, part I, page 452, and in Taylor I, 71,
no. 1987. It consists of 12 Cadjan (palm) leaves of the
size 12¼" by 1⅜". Each page contains 8 lines. It is
written in Grantha character. Neither the date of the
manuscript nor the name of the copyist is given. Many
leaves are missing and more are so injured that it is
impossible to make out the contents. The last page
contains in the middle the colophon of Chapter XLVII.
The comparison has to be given up after a few lines of
Chapter XLVI (Toraṇa).

xvi PREFACE
(K) These two manuscripts belong to the Palace Library of
Trivandrum. They are referred to in Aufrecht part I,
page 452, and described in Oppert II, 532, 6125, 6976.
Of these, 532 and 6976 are not available and seem to
have been lost. Of the two Trivandrum manuscripts,
one is 6125 and the other is a new one, in the sense that
it is not mentioned in any catalogue. Aufrecht thinks
(page 452) that one of these three described by Oppert
was "quoted by Rām Rāj." But there seems no
reason for Aufrecht's supposition. From the descrip-
tion it seems on the contrary that Rām Rāj referred to
the one called here 'J.'
Both of the Trivandrum manuscripts are written on palm-
leaves in Malayalam character. One of them contains 63 ślokas on
25 folios with 9 lines to a page of 7½" by 1¼" and goes as far as the
first 55 verses, with a Malayalam gloss. The other contains 6 folios
of 7½" by 1¼" with 9 lines to a page. This second one is the
gloss and omits the verses 5, 9, 10, 14, 24, 28, 38, 40, 45, 48,
52 and 55. They deal with Śaṅkhu-sthāpana and the illustration
of the Āyādi-shaḍ-varga; but comparison and collation are not
possible. Apparently they are not copies of the Mānasāra and
seem to belong to a different work. They are retained in
Appendix ग.
Aufrecht, part III, page 97, refers to the Reports on Sanskrit
Manuscripts in Southern India by E. Hultzsch, vol. II, page 37,
no. 1103. These manuscripts are the same as Burnell, Tanjore,
page 62, nos. 11080 and 11081, and identical with 'F' (11081)
and 'G' (11089).
Aufrecht, part III, page 97, refers to Hara Prasad, I and II,
Calcutta, 1898, 8, II, no. 161. It is, perhaps, the same fragmentary
manuscript, which now belongs to Mr. Nagendra Nath Basu, and
appears to be a recent copy of one of the manuscripts described
above.

PREFACE xvii
Of all these manuscripts, 'A' is the shortest. The characters
are legible, clear and bold, but the language is very corrupt. The
corruptions are not due only to the copyist, but are to be ascribed
to the sources from which 'A' was copied. It does not differ
materially from the first and second parts of 'F.' In fact 'A' and
'F' resemble each other so closely as to suggest a common origin,
and may be taken to form one group.
'B' is the most incorrect manuscript; but, such as it is, it
does not differ from 'D.' The various corruptions, except those
which are obviously copyist's conjectures and variae lectiones, are the
same. I am inclined to think that 'B' had already been copied from
'D' when the latter was removed to the India. Office. At any
rate their common origin is apparent. These two will thus form
one group.
'C' is also interpolated and stands aloof, although it shows
many striking coincidences with such parts of 'J' as do not agree
with 'I.' Eggeling says (Catalogue, page 1154) "the volume ('C')
was copied from two different manuscripts . . this would account
for certain repetitions, as well as for certain orthographic peculiarities
(such as yeka for eka, vuttara for uttara), pointing to a Telegu source."
That the original of 'C' was very damaged is clear from such marginal
remarks of the copyist as "Asmin sthale patrāṇi navāṇi nyūnāni"
(here are lost nine leaves).
'G' and 'H,' each of which has the same fifteen chapters only,
are in a fairly good condition. They are copies from the same
original, differing just enough to show that they have undergone
the usual corruptions, which a long course of copying and re-copying
under different circumstances renders unavoidable. They may,
therefore, be classed under one group.
'I,' which is the Codex Archetypus of the present edition is the
most perfect manuscript available of the Mānasāra. It is likely
that 'I' was copied from 'J' when the latter was in its complete
form. This would account for the displacement of some lines in

xviii PREFACE
the same chapter, as well as in different chapters, through copyist's
blunders or want of revision. The lacuna of 'I' may thus be
ascribed to negligence in reading the extremely injured parts of 'J,'
since we have such obvious common mistakes as varṇayet for
van-nayat, etc. They may be taken to form one group.
With regard to the provenance of the Manuscripts it is
unfortunately impossible to give much definite information.
'A,' 'C,' 'D,' 'E,' belong to the India Office. 'B' comes from Cal-
cutta and 'H' from Bombay. 'F,' 'G,' 'I,' 'J' come from Madras.
The Telegu character of 'D' and the Grantha and Tamil characters
of 'F' and the Grantha character of 'G,' 'I' and 'J' point to
their South Indian origin. "B' seems to have been written in
Bengal and 'H' in Northern India. In spite of their Devanāgarī
character, certain orthographic peculiarities, such as yiśa for īśa and
yeka for eka, 'A,' 'C' seem to be copies of South Indian originals;
or more probably, they were transcribed by South Indian copyists.
The critical apparatus furnishes abundant warrant for the
division of the manuscripts into such groups, and it could easily be
supplemented by a number of details, omission or addition, punc-
tuation, particular forms of letters, etc., which are not of sufficiently
noteworthy importance. Within each group the manuscripts are in
close relationship. This is proved by a number of variants common
to each group and especially by long lacunæ.
The relationship of the manuscripts within the above mentioned
families ('A,' 'F'; 'B,' 'D'; 'C'; 'G,' 'H'; 'I,' 'J') is a
matter of considerable difficulty. The nature of the work renders
it probable that the relationship should vary in various particulars,
and to some extent this expectation has been justified.
Of the peculiarities of the groups the following may be mentioned:
'A,' 'F' regularly write Kamṭha, Kamdhara, Kampa-madhyena; 'B,'
'D,' and 'I,' 'J,' write the same words as Karṇa, Kaṁdhara, Kam-
pamardhena. 'B' and 'C' have sometimes yeka for eka, yiśa for
īśa, etc. Such peculiarities as are common to almost all the manus-
cripts will be found treated more elaborately under the Critical Notes.

PREFACE xix
General method followed in preparing the text
After the brief survey of the present condition of the manu-
scripts, it remains for me to indicate shortly the lines which I have
followed in preparing the text of the first edition from the materials
furnished by the codices.
The variations in the manuscripts consisting in spelling, words,
phrases, sentences and grammatical concordance have been recorded.
Errors in metre, when accidental or results of bad copying, can
easily be corrected. But such errors in the Mānasāra are too
numerous and they can hardly be corrected without tampering with
the original. In all sacred works they have been respected by
mediæval Indian scholars, and accounted for as peculiarities of
holy sages (ārsha). The first editor of the Mānasāra has no other
alternative than to take all these errors as ārsha.
I believe I have thus included in the Notes and Appendices
more than what has any real bearing on the text. This principle
is certainly not economical. But this is, under the circumstances
to be noticed below, unavoidable and not altogether useless.
The chief object of critically editing the ancient manuscripts
is to prepare an intelligible text. At the same time no one could
deny the "mischief done by conjectural criticism of classical scholar-
ship." But with a view to preparing an intelligible text of the
Mānasāra, the ungrammatical or "barbarous" style of which will
be found discussed elsewhere,¹ a certain amount of emendation is
indispensable.
When the office of the Editor, Commentator, and Translator is
all combined in one, the following principle, which, if not sound, is
at least safe under the peculiar circumstances, has been followed in
this edition :
"Let an editor give what there is, and let the commentator
say what might be and what ought to be."
¹ Critical Notes, pp. 1-310; Appendix to Indian Architecture, pp. 190-214.

xx PREFACE
This method has the merit of satisfying the two divergent schools
of critics, namely, scholars who are in favour of an intelligent
discrimination between what are different readings and what are
mere blunders due to ignorance of authors or negligence of copyists,
and those who insist upon a faithful reproduction of even the most
obvious and unquestionable mistakes.
Working on this principle, such emendations as have been thought
indispensable for the constitution of an intelligible text have been
invariably marked as such both in the text, within brackets as is
usually done in editing an ancient inscription, and also in the Notes,
and the actual readings of all the manuscripts have been carefully
recorded. But I have kept in view the risks with which any attempt
at emendation is beset in the case of so difficult a text as the Mānasāra
and have accordingly endeavoured to exercise the utmost caution.
For the same reason, I have abstained from suggesting conjectural
restorations for all the passages in which the text, as found in the
Codex Archetypus, showed a lacuna, except where it could be restored
from the nature of the context or by reference to related passages
with a reasonable degree of certainty.
Under the circumstances pointed out in the description of the
existing manuscripts, ‘I’ is the only possible basis of this edition.
From this a departure has been made in favour of ‘J’ only where the
reading of the latter appeared distinctly preferable and to represent
a close approach to the original. In every such case the text of
‘I’ and all other manuscripts has been carefully indicated in the
Notes. This has also been done in every instance where the text
of ‘I’ had for the sake of elucidation to be corrected by any
other manuscripts. Whenever the readings of ‘I’ seemed to be
superior or equal to the variations of ‘J,’ and accordingly were
retained in the text, the readings of ‘J’ have been recorded in the
Notes. All words or letters, which ‘J’ or any other codex has
supplied, where ‘I’ had a lacuna or otherwise, have been distinctly
noted.

?
Wait! Look at the letter after तु:
It is म!
Wait, तु म . ग? No, तुम्ब? तुम्बर?
Wait, what if it's NOT तु?
What if it's तु ङ्ग?
Wait, let's look at:
तुङ्ग? But why is there a comma or dot?
Wait, is it an abbreviation? No, it's:
तुङ्ग for त्रिभाग? No! Look at त्रिभाग:
Wait, is that second word त्रिभाग?
Let's read the characters of the second word:
First: त्रि or वि or श्रि?
It has a loop with a tail! That's वि!
Wait, वि or त्रि?
Look at त्रि: त with r stroke त्र, with i matra त्रि.
Look at वि: round loop on left of vertical bar, with i matra वि.
Wait, look at the third character of that second word:
It's ग or न?
It is न! विमान!
विमान! YES! विमान!
The second word is विमान! for विमान,!
Of course! Vimāna is the central architectural term of Manasara!
Now