12001 automatic labels ←

chunk 7: 12027 length, accent, labels

→ 13012 bendy roots, nadI

12027 What has the duration of u U U3 is short long extralong. UkAlojjhrasvadIrghaplutaH
12028 But only if it's a vowel. acazca
12029 Acute means high tone. uccairudAttaH
12030 Grave means low tone nIcairanudAttaH
12031 Falling is mixed samAhArassvaritaH
12042 A tatpuruSa is a karmadhAraya if its two halves have the same referent as the whole. tatpuruSassamAnAdhikaraNaHkarmadhArayaH
12045 Nounbase is what has meaning, is not a root, and has no affix. arthavadadhAturapratyayaHprAtipadikam
12046 kRt, taddhita, compound too . kRttaddhitasamAsAzca
12047 Shorten neuter nounbase hrasvonapuMsakeprAtipadikasya
12048 Shorten go- and GI-/Ap-enders when they are latter and subordinate. gostriyorupasarjanasya
12051 When there is lup, gender number are like original lupiyuktavadvyaktivacane
13001 roots are bhU etc. bhUvAdayodhAtavaH
13003 Final consonant . halantyam
13004 But the final tu s m of a vibhakti aren't . navibhaktautusmAH
13005 Ji Tu' Du at the start . AdirJiTuDavaH
13006 The initial S of an affix . SaHpratyayasya
13007 cu Tu of affixes . cuTU
13008 l z ku , but not in a taddhita lazakvataddhite
13009 Labels are invisible. tasyalopaH
13010 When a rule has two lists of the same length, understand "respectively". yathAsaGkhyamanudezassamAnAm


U-kAlo 'j jhrasva;dIrgha;plutaH ONPANINI 12027
What has the duration of u U U3 is short long extralong.mmmmmmmmm glosses glosses ^ M- C+ 79

(The "U3" here means the extralong U sound. Extralong vowels are so uncommon in writing that they don't have letters of their own; in the veda manuscripts they are written with the letter for the nearest long, followed by a figure 3.)

Example. In the word gurU "two teachers" the two vowels are said in exactly the same way EXCEPT that the second one lasts longer. The u is said to be a short and the U is said to be a long.

And, of course, an extralong is a vowel longer than a long .

Even though pANini just says that short, long and extralong have different durations, later grammarians came up with the idea that shorts last for a certain time, longs twice that, and extralongs thrice that. It is a good idea to try to follow that rule when chanting verse or the veda. See mAtrA.

See exception acazca below.


acaz ca ONPANINI 12028
But only if it's a vowel.mmmmmmmmm glosses glosses ^ C+ 80

In the rules, the terms " short", " long" and " extralong" mean vowels only.

Example. If any rule says "replace with a short", that rule will replace a long vowel with a short vowel, or will replace a long vowel with itself, but it will never replace a consonant.


uccair udAttaH ONPANINI 12029
Acute means high tone.mmmmmmmmm glosses glosses ^ C+ 81 accent

The vowel that has the acute accent is pronounced with a pitch higher than the others.

Back to accent.

gaN(A)nAM tvA gaN(a)patiM havAmahe kav(i)M kavIn(A)m upam(a)zravastamam

listen to it here --

LISTENTO shri rudram


nIcair anudAttaH ONPANINI 12030
Grave means low tonemmmmmmmmm glosses glosses ^ C+ 82

A grave vowel is pronounced with the average, normal tone. Everything is grave except the acute and the falling.

Back to accent.


samAhAraH svaritaH ONPANINI 12031
Falling is mixedmmmmmmmmm glosses glosses ^ C+ 83

A vowel has a falling (svarita) accent when its tone is mixed of the previous two -- it starts acute and ends up grave.

When a vowel has to be acute according to the rules, ands the next one has to be grave, the grave is automatically replaced with a svarita. So, most of the falling vowels you can hear come right after an acute vowel.

The exception to that is the so-called "independent svarita". In some cases we can hear a svarita right after a grave vowel. This happens, for instance, when an acute (vi) is followed by a grave a, and then they merge into vya. In such a case the resulting ya sound starts high and ends up low.


also here --


the word vyupta has a falling accent on the vyu, even though the ca of kapar(di)neca is grave.

Back to accent.


tatpuruSaH samAnAdhikaraNaH karmadhArayaH ONPANINI 12042
A tatpuruSa is a karmadhAraya if its two halves have the same referent as the whole.mmmmmmmmm glosses glosses ^ M- C+ 84

A tatpuruSa is a compound that means the same thing as its latter.

For instance, a firefighter is a fighter, so the compound "firefighter" is a tatpuruSa.

And a blackbird is a bird, so the compound "blackbird" is a tatpuruSa.

If the first half of the tatpuruSa also means the same thing as the whole and as the second half, then it is a karmadhAraya, in addition to being a tatpuruSa.

A firefighter is a fighter, but he is not fire, so the compound "firefighter" is a tatpuruSa, but not a karmadhAraya.

And a blackbird is a black thing, so the compound "blackbird" is a karmadhAraya, and a tatpuruSa as well.


arthavad a-dhAtur a-pratyayaH prAtipadikam ONPANINI 12045
Nounbase is what has meaning, is not a root, and has no affix.mmmmmmmmm glosses glosses ^ M- C+ 85


azva- is a nounbase according to this rule because it means "horse", has no root, and has no affix.

So is zveta-, as it means "white", has no root, and has no affix


zvet is not a nounbase because it doesn't mean anything

azvas "horse" is not a nounbase because it has an affix ( su)

the word zvetA "white", as in nadI zuklA, is not a nounbase, because it has Ap and su

kSip means "throwing" and has no affix, but it is not a nounbase because it is a root

Forms like saurya- "solar", that have an affix inside (here aN), are not made into nounbases by this rule, but they are nounbases anyway because the kRtta right below.


kRt;taddhita-samAsAz ca ONPANINI 12046
kRt, taddhita, compound too (are nounbases).mmmmmmmmm glosses glosses ^ M+ C+ 86

Rule arthavada above says that some words are nounbases. This rule clarifies that --

(1) kRt-enders are nounbases

Example. gup "lead" is a root, and kta is a kRt affix. Joining those we get gupta- "hidden", which is a nounbase.

(2) taddhita-enders are nounbases

Example. sUryaH "the Sun" is a noun, and aN is a taddhita affix. Joining those we get saurya- "solar", which is a nounbase.

(3) compounds are nounbases

This "compounds are nounbases" thing in fact means --

"when we remove the sup and feminine affix at the end of a compound, what is left is a nounbase".


Joining rAjJas and puruSas we get rAjapuruSas, which is a compound.

Removing the sup at the end of rAjapuruSas we get rAjapuruSa-, which, according to this rule, is a nounbase.

Joining sarveSAm and bhUtAnAm we get sarvabhUtAnAm, which is a compound.

Removing the sup at the end of sarvabhUtAnAm we get sarvabhUta-, which, according to this rule, is a nounbase.


hrasvo napuMsake prAtipadikasya ONPANINI 12047
Shorten neuter nounbasemmmmmmmmm glosses glosses ^ M+ C+ 87

Not many neuter nounbases end in a long vowel. So my examples are compounds of rai-, nau-, rootnoun --

atirai- + su + kula- + su → atiri kulam "super-rich family extravagant family"

atinau- + su + kulam → atinu + s + kulam svamorna atinu kulam "disembarked family, the family that came out of the ship"

bahu- + gramaNI- + su + kulam → bahugrAmaNi kulam "family that has many mayors"

atirai- + au + kule → atiri + au + kuleatiri + zI + kule ikocivibhaktau atirinI kule Natvam atiriNI kule "two super-rich families"


go;striyor upasarjanasya ONPANINI 12048
Shorten go- and GI- Ap-enders when they are latter and subordinate.mmmmmmmmm glosses glosses ^ M+ C+ 88

Examples --

citra- + go- → citragu- "Mr Spotty Cows"

nis + kauzambI- → niSkauzambi- "that left kauzambI"

ati + khaTvA- → atikhaTva- "that no longer needs a bed"

The su added after these GI Ap will not be deleted by halGyA --

atikhaTva + suatikhaTvas "he no longer needs a bed, he's an ascetic"


lupi yuktavad vyakti;vacane ONPANINI 12051
When there is lup, gender number are like originalmmmmmmmmm glosses glosses ^ C+ 89

Making a long story short, this rule allows sentences like "I'm going to the Aussies" to mean "I'm going to the country where the Aussies live" and still be grammatical.

An example. The kurus are a people, descendants of a certain kuru. The country where the kurus live is called "the kurus", used like this --

kurUn gacchAmi "I'm going to the kurus"

This kurUn word here means a certain country so called, it does not mean its inhabitants. When used in this sense, the word is breaking the ordinary rules of grammar. For instance, rule dvyeka says that a noun that means ONE country must be singular, and some other rule says that when we name a country after its inhabitants, we must add some suffix after the inhabitants' name, such as, for instance, aN.

To bring this kurUn word, which is masculine and singular, into the shed with the other country names, grammarians say that it is formed like this: starting at kurUNAM rAjyam "the kingdom of the kurus", we first replace it with kurUNAm + aN, then replace the aN with lup. This makes the nounbase kuru-, which (like the original word kuruNAm) is masculine and takes plural affixes, but means one country anyway. If we had used luk instead of lup to delete the aN, other rules would have made this nounbase into a neuter singular.

.. add the KAZIKA and more examples of masc. pl country names


bhUvAdayo dhAtavaH ONPANINI 13001
roots are bhU etc.mmmmmmmmm glosses glosses ^ M- C+ 90

roots are things invented by grammarians that appear in a list called the dhAtupATha. Such as




Roots cannot be used alone. Before use, we must add to them either a tense (to make a verb), or a kRt affix (to make a nounbase).

The tense will always be replaced with a tiG, by lasya --

kSip ( a root ) + laT ( a tense ) lasya kSip + tipksip + za + tikSipati "throws" ( verb )

The nounbase will always be affected by GyApprA --

kSip + tRc ( another kRt affix ) + su puganta kSeptR- ( nounbase ) + su Rduzana kSeptA "thrower" ( noun )

kSip ( root ) + kta ( a kRt affix ) + aukSipta- ( nounbase ) + au ( a sup affix ) → kSiptau ( noun ) "both were thrown"

In some rules, the word dhAtu " root" means rootnoun. One such rule is atvasa.


hal antyam ONPANINI 13003
Final consonant (in grammar list is label).mmmmmmmmm glosses glosses ^ M- C+ 91


In the zivasUtra list, a i uN R Lk e oG ai auc etc,

the N k G c are labels, because they are added at the end of u, L, etc.

In the list tiptas (tip tas jhi sip thas tha... )

the p of tip sip mip, and the T of iT', are labels. The real affixes are ti si mi, i

the s of tas thas.. are not labels because of the exception navibha.

in the dhAtupATha (list of roots), (bhU edha spardha... pluG... )

the G of pluG is a label, the real root is plu.

and the edh root had an a label added just so that this rule does not make dh into a label

in list svaujas (su au jas am auT zas TA bhyAm bhiS...)

the first T is a label, the real affix is just au


na vibhaktau tu;s;mAH ONPANINI 13004
But the final tu s m of a vibhakti aren't (labels).mmmmmmmmm glosses glosses ^ C+ 92


in the list tiptas (tip tas jhi... ta AtAm jha... ) the s of tas and the m of AtAm are real

in the list svaujas (su au jas... Gas os Am...) the s of jas Gas os are real, and so is the m of Am.


Adir Ji;Tu;DavaH ONPANINI 13005
Ji Tu' Du at the start (are labels).mmmmmmmmm glosses glosses ^ C+ 93

The Ji Tu' Du at the start of a root in the dhAtupATha is a label. Such roots are called JIt, Tvit, Dvit roots. For instance, the root kR "make" is listed in the dhAtupATha as (Du)kR(J), so it's a Dvit root.

See --

Ji for examples of JIt roots

Tu' for examples of Tvit roots

Du for examples of Dvit roots


SaH pratyayasya ONPANINI 13006
The initial S of an affix (is a label).mmmmmmmmm glosses glosses ^ C+ 94

So the S of Svun is a label. But not the non-initial S of the affix TiSac of mahiSa- "buffalo". And not the initial S of the non-affix SaS- "six".

Whatever has S label is a Sit.


cu;TU ONPANINI 13007
cu Tu of affixes (are labels).mmmmmmmmm glosses glosses ^ M- C+ 95

Examples of cu labels --

c of Tac

j of jas

Examples of Tu labels --

T of Tac

N of Nya


la;za;kv ataddhite ONPANINI 13008
l z ku (in affixes are labels), but not in a taddhitammmmmmmmm glosses glosses ^ C+ 96

Examples: G of Gas, k of kta, z of zas, z of zap, l of lyap are labels.

However, the k of kan kAmyac, and the z of zas', are real, because these are taddhita affixes.


tasya lopaH ONPANINI 13009
Labels are invisible.mmmmmmmmm glosses glosses ^ M- C+ 97

Talkaround. The label letters (a.k.a. it letters), namely, those described in the previous rules, must often be ignored.

In particular --

(1) A label never makes it into an actual word.

So, when we join nRtI~ and zyan and tip, to make a word that means "is dancing", we don't get nRtI~zyantip. The letters I~, z, n and p are just labels, so the actual word is nRtyati --

devI nRtyati jyotsnAyAm "the queen is dancing in the moonlight"

(2) In any rule, "first letter of" means "first non- label letter of".

Example. When we join pac + kta, rule pacovaH says: "replace the first letter of (k)ta with v". This "first letter of kta" means "the first non-label letter', namely the t. So kta turns into kva, not vta --

pakvA nRpeNa kukkuTI "the king cooked a hen"

(3) When counting letters, labels aren't counted.

Rule halGyA says "delete sip if it has one letter". si(p) ordinarily has two REAL letters, the s and the i; the p is a label and doesn't count. After itazca works, si(p) turns into s(p), which has ONE real letter. And then halGyA can work.


yathA-saGkhyam anudezaH samAnAm ONPANINI 13010
When a rule has two lists of the same length, understand " respectively".mmmmmmmmm glosses glosses ^ C+ 98

Exception to sthAnentaratamaH.


TAGasi rule replaces TA Gasi Gas with ina At sya' respectively.

ecoya rule replaces e oG ai auc with ay av Ay Av respectively.

tas;thas;tha;mipAM tAM;taM;t%AmaH replaces tas thas tha mip with tAm tam ta am respectively.

12001 automatic labels ←

chunk 7: 12027 length, accent, labels

→ 13012 bendy roots, nadI