chunk 72: footnotes 3 chunk 72: footnotes 3
about
verb classes
unchanging has neither
non-third person
replacing a thing with itself looks pointless
how to learn rule
about the most common sandhis
key to exercises of sandhi four
about
Which roots get
About the
auxiliary roots
About accents.
rules that delete root nasal
the auxiliary
prohibition
Metri causa means "to make the verse fit".
epics have oddities.
about hidden words
Accent rules are only applied when reciting the veda.
Extralong is a vowel longer than a long.
No
about
about
About
formation of
Spelling of
About hangers.
types of affixes
split
about
rules that delete the calling
MCM, sandhi of
Accurate translation of
replace with
about the
about the ages of the world
the six duties of a brahmin
Warnings about weird spellings in this site.
Rule
Cluster lengthening after short.
Uses of the second endings.
warning about flat bent rules
root changes before
Optionality of
soft
about strange symbols in rules
what are the
rules that form the flat soft
bent soft
^ 
There are two groups of vibhakti affixes --
The prAgdizIya affixes are vibhakti affixes, by rule prAgdizovibhaktiH
The sup and tiG affixes are vibhakti too, by rule vibhaktizca.
Several rules work on vibhakti affixes, such as aSTana::A, tyadAdInAmaH, navibhaktautusmAH.
^ 
The roots listed in the dhAtupATha are divided into ten groups called verb classes. These are the numbers of the ten verb classes, the affix that their roots get (when before a hard doer affix), my name for the class, and the rule that adds the affix --
[1] zap zapclass kartarizap
[2] luk lukclass adiprabhRtibhyazzapaH
[3] zlu zluclass juhotyAdibhyazzluH
[4] zyan zyanclass divAdibhyazzyan
[5] znu znuclass svAdibhyazznuH
[6] za zaclass tudAdibhyazza
[7] znam znamclass rudhAdibhyazznam
[9] znA znaclass kryAdibhyazznA
These ten affixes are called vikaraNa ("changer") affixes by grammarians other than
Rules one to nine above only work when the root is right before a hard doer affix. Rule ten, however, ALWAYS adds Nic, no matter what is afterwards. If what is afterwards is a hard doer affix, the Nic becomes
The lukclass roots do not get a luk affix; I just use the word luk in the list to mean that they get no affix at all.
The numbers [1] [2] [3] etc appear in inria reader when you type a verb of that class that contains the affix. If inria shows " ca", the verb has causative Nic and works as if it were a [10]. See Using parsers to split verbs into root plus affix for examples.
Summarizing --
" zap luk zlu zyan znu za and znam u znA Nic are the ten affixes"
^ 
When I say " noun", I mean
Some examples of nouns, with the sup they have. What is before the sup is either a nounbase, or a nounbase with a feminine affix added at the end.
Notice that you cannot trust the translation of a sentence to know if something is a noun or not. In
^
Whatever has a tiG at the end is a verb (in Sanskrit:
as + laG →
Students should keep in mind that whatever does NOT have a tiG at the end is not a verb, no matter if it translates into English as an English verb or not. As an example, in this sentence --
there are three words that have a root inside and mean an action, but none of them is a verb, even though their English translations may be English verbs ("went", "saw" are verbs). That's why when you drop that sentence into inria you get no red words. The red color means that there is a tiG at the end.
inria reader always paints red whatever has a tiG at the end, and paints blue whatever has a sup. The exception to that is the Am'' verbs like
^
An old jingle goes,
Sanskrit words can be divided in three groups, that inria paints blue, red and pink --
blue -- nouns OR
red -- verbs OR
pink -- unchanging OR
blue are nouns, made by adding a sup ( noun ending) after a nounbase. We may add different sup to the same nounbase.
red are verbs, made by adding a tiG ( verb ending) after a root. We may add different tiG to the same root.
Everything else is pink -- unchanging.
Pink words either don't get a sup tiG at all, or always get the same sup. So they always look the same.
Feel free to call these unchanging words with other grammatical terms that will help you remember. For instance you might call
Notice that ktvA and tumun are kRt affixes, so in a sense,
(nonthirdpersonliTisver) (non)
^ 
The liT tense has been seldom used since the epics. And even in the epics, only the third person endings are found -- Nal atus us ez AtAm irec.
As all those start with vowels, they never get iT. Which means you don't need to pay much attention to the rules, such as kR;sR;bhR;vR;stu;dru;sru;zru and others, that add iT to the valAdi liT (namely thal va ma se''' dhve''' vahe''' mahe'''),
If we take seriously the
And similarly, you can use the first person liT only when you talk about something that you don't remember having done or claim to not have done --
Frankly -- If you are going to use a first or second person liT, do not worry about the
(replacingathingwithits) (replacit)
^
Rule atom says "replace su with am and am with am". At first sight, replacing one thing with itself looksstupid, but lets look at the previous rule too. Rule
svamorna says "after neuter, replace su with luk and am with luk"
and the next rule is an exception to it,
atom , and it says "but if the neuter ends in
Therefore, when su follows a neuter, we replace it with luk, UNLESS the neuter ends in
And, when am follows a neuter, we replace it with luk, UNLESS the neuter ends in
In English, we might have worded the second rule as --
"but if the neuter ends in
However, in rule jargon, it turns out that saying ' replace su with am and do not replace am ' is a waste of words, because saying "replace su and am with am" is much shorter, and the students do not have to spend so much time chanting the rules.
^
Because of rule aTku, when we add
So we have to figure out if we use that rule or not whenever we add any affix that contains a
In my teaching experience, every student takes a long time to get used to this, and this causes a lot of pain, because when you mistype
Yet if you had learned Sanskrit at home, by imitating what you hear, you would say automatically
Thing is, there is a subtle difference between the way to say the tap-R sound in Spanish or English and the way that is natural in Sanskrit.
(1) When you say "diabetes" in most American accents, after saying the
(2) Yet in good Sanskrit, after say the tap
In ancient Sanskrit, all of the letters
When trying to say
When trying to say
So if you are used to moving your tongue correctly, rule aTku can be translated this way --
"
and your tongue is slightly raised only when you raised it at
This means that the way to learn to apply this rule while talking is learning to move your tongue correctly. If when you try to say
Some words for practice --
(aboutthemostcommonsand) (mos)
^ 
To find out which are the most common sounds found at the start and ewnd of words in Sanskrit, I undid the sound changes in a dozen thousand zlokas lines of the mahAbhArata, this way --
and counted how many times each letter appears at the beginning and the end of words. My numbers show that, in my sample --
Roughly 40% of words start with a haz letter
another 40% start with a khar letter
the other 20% are vowels, half of which are
As for the ends of words,
55% of words end in a vowel (of those, about half end in
18% end in
13% end in
3% end in
4% end in an
4% end in
4% end in
and the remaining ways a word can end (namely
This gives us a rough idea of which sound change rules we have to learn first. The most important rule is "no change when vowel goes before consonant". As 0.4 multiplied by 0.55 is 0.44, which means that if we take two words at random from the mahAbhArata, we have a 44% chance that the first ends in a vowel and the second starts with a consonant, so no change is needed. The next most commonly used rule is "
vowel + consonant 44% - no change
m + consonant 15% -
as + haz 5% -
as + khar 5% - different results
s + consonant 3.2% - different results
d + consonant 3.2% - different results
n + consonant 3.2% - different results
As + khar 1.2% - different results
As + haz 1.2% -
as + a 1.3% - combine both into
As the rules that do not give different results are easier, we must start with --
vowel + consonant 44% - no change
m + consonant 15% -
as + haz 5% -
As + haz 1.2% -
as + a 1.3% - combine both into
^
These are the answers to the exercises in examples of sandhi four .
^ 
zrad is not a preverb, but it is a gati before dhA "put", making a verb
^ 
These are the rules that turn
Rules kAs;pratyayAdAmamantreliTi ff
explain which roots get Am'' instead of liTidhA.
Rule AmaH
deletes the liT affix.
adds that affix to an auxiliary then glues them after the Am''.
And rule Ampratyaya clarifies that two of the auxiliaries never get bent affixes.
^
Rules kAspra ff explain exactly which roots get Am'' before liT. But summarizing those rules here can do no harm.
The rule-of thumb is that roots with one vowel get liTidhA, like
and the rest get Am'', like
(In those examples, the Asa is added after Am'' by rule kRJcA).
Now the exceptions to the general rule.
(1)
(2) So do the one-vowelers that start with a heavy vowel that is not
(3) And these roots make their liT both ways --
(4) UrNu has two vowels, but gets no Am'' --
Note. UrNu did not reduplicate into
^
Short explanation: the uNAdi affixes are affixes that are not mentioned in the
Now the long explanation.
Many Sanskrit nounbases are formed by adding kRt affixes after roots. For instance,
Most of the affixes mentioned in the rules are quite useful, in the sense that what they can help to explain the meanings of many words. For instance, rule NvultRcau tells us that all roots can get Nvul, from which we deduce that we may say
Others are less useful, like ktin, that is added only to some roots, or snaJ, that is added to
The uNAdi affixes are affixes invented by other grammarians before and after
As you can use a dictionary to check that
^ 
The auxiliary roots are kR, bhU and as.
Sometimes they are used to make auxiliary verbs such as
These verbs are called auxiliary verbs when they are used as if they were affixes. This only happens after Am'' and cvi.
Example after
Example after
The forms made from as + liT, such as
^
In the Sanskrit described by
The other vowels are called
In
No one knows why, or when, the Sanskrit teachers not involved with the veda stopped teaching accents. My guess is simple: laziness. Half of the rules of the grammar are related to accent, so if I put up with my students using wrong accents, I not only cut my work in half, I can also be happily ignorant of half the rules.
Example. In kosher Sanskrit, these two are different words pronounced differently --
but I pronounce both words the exact same way, and teach my students to do the same. This way I don't have to bother to tell them that
Some people will tell you that it is forbidden to use accent when not reciting the veda. They lie thru their teeth. No
From now on, when anyone says
(rulesthatdeleterootnas) (rulesthn)
^
Rule that erases the final nasal of a root --
anudAttopadeza (makes
Rules that erase the nexttolast nasal of a halanta root, and exceptions --
64023 znAnnalopaH after znam
64024 an-iditAM hala:: upadhAyAH kGiti before kGit
64025 daMza;saJja;svaJjAM zapi before zap
64027 ghaJi ca bhAva-karaNayoH
^
By asterbhUH, the root as turns into bhU before liT, so that as + liT jhi makes
Yet this only happens when as + liT is used as an ordinary verb. When it is an auxiliary, that is to say, when it is added after Am'', it stays as, and then we get --
As in --
^
The particle mAG is used to forbid or dissuade. It can be used with a loT, where
yet instead of the loT, the luG can be used. This luG will not get aT AT, because of namAGyoge --
if there is both mAG and sma, the laG can be used, again with namAGyoge --
By alaGkhalvoH, when the sense is "stop doing", we can use
We can also use a tumun, or an action noun with TA --
^
"Metri causa" (say that as
This expression is used when Latin and Ancient Greek poets break some grammar rules to make the verse fit.
Latinists take good note: in the Sanskrit tradition, good poets are supposed to never breach
Yet, the epics appear to be older than
"wealth comes from
"
According to good grammar,
Scholiasts of the epics will excuse ungrammaticalities like
coulson says that the makers of the epics composed Sanskrit "intuitively" with no knowledge of
^
The word ArSam means "belonging to the old language of the
Example. Suppose a zloka line has 17 vowels, such as
which boils down to saying "the extra syllable is not a typo".
Commentators do not always use that word. For instance, this line is a clear violation of rule samAse '-naJ-pUrve ktvo lyap --
But the commentary might just have
^
Suppose I ask "where did the squirrel go?" and I say "under the tree". If I say that, maybe my English grammar is not very good, but you will understand that I mean "the squirrel went under the tree", "the squirrel is under the tree", or "look under the tree".
If you are asked "did John catch the squirrel", you won't say "John caught the squirrel". The normal thing to do is skipping the words that will be easily understood even if missing. So, you might answer "he caught it", or "Paul caught it", or "he caught the moose", or just "yes".
English grammar is very picky about what happens when a word is hidden. I'm told that "under the tree" is bad grammar, and I must say "it's under the tree" -- replacing "squirrel" with "it". Sanskrit grammar, however, is not so picky. Same as in Japanese grammar, if a word is necessary you say it, and otherwise just don't say it. If you say this --
everybody will understand that you mean "by
The
In those rules, the word
(accentrulesareonlyappl) (acce)
^ 
Nowadays, the Sanskrit accents prescribed by
But no other Sanskrit teacher cares about teaching accent. This is sort of a mystery, because it is abundantly clear that
Now, how did we come to this situation? Personaly I blame
I'll do a prophecy. I guess that, as
^
The extralong vowels, also called
Ther come in two flavours.
First, the extralong vowels of the veda -- some vowels must be chanted with a duration of three mAtrA. You know which because they are written as a long vowel, but with a figure 3 after them.
And then, those mentioned by
Some rules that mention the extralong --
When you call someone from far away, rule dUrAddhUteca says, you may make the last vowel last more than a long.
Rule kSiyA says that you can use sometimes an extralong when you are angry at rudeness, blessing someone, or giving several commands.
^
The first half of a compound is always a word (by supodhA). Now, in some compounds, like
There are dozens of rules allowing exceptions to the same-word like this one, but you don't need to worry about them much, because they usually happen in compounds, such as
These changes happen because fluent speakers will automatically apply raSA to whatever "feels" as if it were a single word. And expressions feel like a single word when they are used a often. So
As words evolve from a compound to a commonly used word, there is no way to tell exactly when a compound will start to feel like a single word. The
^
raSA only affects the
So, there is no
But there is Natvam of the
or before a funny --
Yet, there is no rule stating " No
Rule raSA is placed after nazcA and before anusvA, and cannot work on M, so it only works on the
^
The affix
zru "hear" + kta "-ed" →
The kta nounbases can be used either as verblikes, meaning "was led, was heard, was thrown" --
or as adjectives, meaning "led, heard, thrown" --
In the feminine, kta gets Ap, by ajAdyataSTAp --
^
Example.
The
Adding kta we get
meaning "was thrown" (when used alone)
or "that was thrown" (when used as an adjective).
Adding ktavatu we get
meaning "threw" (when used alone)
or "that had thrown" (when used as an adjective, which seldom happens).
ktavatu has label
ktavatu has label
The feminine gets GI, by ugitazca
Example sentences --
^
Some rules have unexpected things, and it is suspected that
Example. According to rule GamohrasvA, we have to double wordfinal
Yet, in sutra iko yaN aci, we have a wordfinal
One way to explain that would be saying that
Yet
So if it is not a bug, it must be a feature. We must conclude that
In this case, the best guess of the experts is that
According to this teaching, you may, if you want, chant sUtra ikoyaNaci as
(formationoflRG) (formationlR)
^
The lRG tense gets sya, as if it were lRT.
It also gets luGlaG itazca tasthastha nityaGGitaH, as if it were laG.
Example --
As in --
(spellingofobeforedelet) (spellinl)
^
^
Rule aNuditsa teaches that the abbreviations ku, cu, Tu, tu, pu stand for these five groups of letters ---
ku =
cu =
Tu =
tu =
pu =
All grammarians after
respectively. So they say
Example. The kAzikA commentary under rule coHkuH "replace cu with ku" explains --
instead of saying
^
Some short words, like ca vA eva iva, are called hangers, because they are always attached after another word. Therefore the expressions --
are always pronounced as if they were single words. Even though they are made of a normal word plus a hanger.
You may write a space before the hanger --
In grammars, hangers are called "enclitics", which is Ancient Greek for "leaners".
Rule anudAttaM sarvam a-pAd'-Adau teaches that hangers cannot start a sentence, verse or half-verse -- they are always hanging after something else.
^
Some groups of afixes --
sup noun endings -- added after nounbases. They make nouns. Described under GyApprA.
kRt affixes -- added after roots. They make nounbases. Described under kRdatiG
tenses -- added after roots. They make verbs.
tiG verb endings -- these replace tenses.
taddhita affixes -- added after nouns. They make nounbases. Described under taddhitAH.
samAsAnta affixes -- added after a compound.
sanAdi affixes.
Some of these are added after a noun to make a root --
and the rest are added after a root to make another root --
^
According to
kR + luT → kR + mip → kR + tAsmi →
while
kR + tRc m + su →
We may also say
So, the grammar forbids using the two words
Yet, in actual use, you will find often two words
"I will give a hundred Gandharva horses to each of you five brothers"
According to
I'll say it again: in theory you may only say
and the bad spelling
used in that same sense.
I call those expressions "split luT", because they are as commonly found as the "split infinitive" is in English, which is also incorrect in theory but commonly used in practice.
Another example --
^
The affix
tarap is much more common that iSThan, but iSThan appears after some very common nounbases.
Example of iSThan --
See also priya;sthira.
(rulesthatdeletethecall) (rulesthl)
^
The
After consonants, Ap, GI by halGyA and eGhrasvAt
After eG and short by eGhrasvAt
After feminine
Therefore, the
Those nounbases are few and far between. They are the ones that end in diphthongs, and the ones that end in a rootnoun from a root that ends in
Examples:
The nounbase
and
^
The words that end in
Turn
Turn
Lose the
The words that end in
Turn
Other words that end in
Turn
All words that end in
Turn
Turn
Turn
Turn
Turn
(accuratetranslationofn) (accu)
^ 
Rule nazchavya, most of the time, changes a wordfinal
(1) The rule won't work on the
(2) Nor on the
(3) Whenever the rule works, we may, if we want, use the atrAnu option and replace with
This
Incidentally. As of 2023, the indology file of the sixth
Back to nazchavya.
^
When a rule says that something must be replaced with guNa, then --
As far as I know no rule ever replaces an
^
The
Some rules that mention aJ --
or aJ --
or aJ x -- aJ comes after
dvaipa-vaiyAghrAd aJ -- dvaipa- and vaiyAghra- have aJ.
prA NirajatA dibhyo 'J -- aJ after living beings and the rajata-class
^
When the nounbases that end in vasu or kvasu are wimpy, the
When they are before strong,
Otherwise, the
The kvasu affix is basically the same affix as this vasu, but it replaces the flat liT affixes. Therefore, it causes reduplication, by liTidhA --
I have not explained how to form all kvasu, but that's okay because they are rare. They are easy to recognize because they have reduplication and
^
The old Indian legends say that time is divided in
The names of the yugas are --
These were named after the four sides of some Indian dice, which, from best to worse, have four dots, three, two and one. The story goes that in
Of course now we are in
So things start very well at creation, and progressively get worse and worse, until they are so bad at the end of
If you see a picture of four cows, the first one standing on four feet, the next one in three, then on two and then on one, now you know what the picture means.
See Wikipedia on yuga for details.
^
The word SaTkarmANi means "the six works" and may mean several list of six activities, like the six duties of a brahmin, the six ascetic practices of yoga, and so on.
The six duties of a brahmin are --
The
(warningsaboutweirdspel) (warnings)
^
1.
If I write
2.
If I write
3.
If I write an isolated Sanskrit word with an
"The word
"Sanskrit for 'again' is
then you must pronounce the
^
This exception can be reworded roughly as --
"When an affix is replaced with another affix, the replaced affix does not count as if it had the same first letter or last letter or number of letters as the original."
Example. When we add laG mip after dviS , we get --
dviS +
Here
dviS +
But when we add laG mip after pac , we get --
paca +
Here someone might say " mip starts with
If the sthAnivad rule just mean "replacement is like original", that reasoning would be correct. Yet, the rule has the word analvidhau at the end. That word means that sthAnivad does not work as far as rules such as atodIr are involved.
In simpler words, when we replace
paca +
(clusterlengtheningafte) (clusters)
^
An old custom forces us to make certain clusters slightly longer when they follow a short. This happens both when chanting, and when talking.
For instance, the words that we write as
will be almost invariably pronounced
Even though, according to
Nowadays, it is better to think that the true pronunciation is
Judging from the manuscripts, this has been done for centuries. In the old manuscripts we find spellings such as
You might ask why, if we always pronounce
(A) According to
(B) According to
(usesofthesecondendings) (usess)
^
A nounbase gets a second ending when --
(1) It is the object of a verb or verblike that means the doer. See karmaNidvi for examples.
(2) When the word is linked to
(3) It is an adjective used as an adverb. In this case the neuter nounbase will be used --
The adverbial usage is seldom in the epics, as, in that sense, the adjective will more often than not agree with the fast thing, instead of becoming neuter and getting second --
(warningaboutflatbentru) (warningr)
^
Even though
The bright side of this is that you can be a little lazy here. If you accidentally say
(rootchangesbeforeyakan) (roo)
^
Roots undergo the following before yak and before soft liG --
Stretchable roots get stretched
vac + yak + te''' →
svap + bhAve loT ta →
but to
but to
Most roots with a nexttolast nasal lose it by aniditA when before yak --
but
^
Rule vAzari says that we may say either
with an H sound first (a visarga sound) and then an
with an
Yet, in my classroom I tell my students that even though
Why do I teach this way?
...
obsolete, see soft
^
The soft liG, nowadays, is used once in a blue moon when it is flat, and never when it is bent.
You never need to use it, because it can only be used to express wishes, and there are many other ways to do that. For instance, using a hard liG works fine. Or a loT.
In spite of that, you may use the soft liG in the laukika if you feel like it, as
So, ordinary students of the language, like you and me, can happily ignore all rules related to the soft liG. Yet, a written test in any Sanskrit univerity can ask you to translate made-up examples like
The flat soft liG makes occasional cameos in the epics. For instance, it exactly appears once in the bhg, at chapter two stanza eight, where the soft
Occasionally you will also find
In this example,
^
The rule ekAco dve prathamasya is spelled as ek%Aco dve. The
List of symbols indicating splits --
(whatarethepANinirules) (whatp)
^
First let me explain what the rules are not. They are not a tutorial to teach Sanskrit language to people that didnt speak Sanskrit. We know that because they are built in such a way that only makes sense to people that are already fluent in Sanskrit.
The rules are basically a manual for teachers of Sanskrit. It is the "book of the teacher". They were not made so that the students could study them, but rather they are an attempt to make all Sanskrit teachers agree so that they all teach the same thing, but not necessarily in the same way.
The rules explain the correct way of speaking and are completely unconcerned about spelling.
(rulesthatformtheflatso) (rulesthft)
^
When they replace soft liG, they are changed by the same rules that make the flat
And then rule yAsuTpa adds
Doing some again, step by step --
Back to flat soft
^
When these replace soft liG, rules jhasyaran and iTot replace jha iT' with ran a', and the four that have
Then, rule liGaHsIyuT adds
sISThAs''' sIyAsthAm''' sIdhvam'''
( sIdhvam''' will sometimes become
Here is sISTa''' done step by step --
soft liG ta →
These affixes have no
Again: do not lose sleep about these. The bent soft liG is so uncommon that I haven't found any example on the wild so far, and I've had to make up my own examples, like the old Vulcan saying --
(No matter how uncommon this tense is in practice, it is kosher because
Back to bent soft
chunk 72: footnotes 3