Gana dosha: Rakshasa is a category, not a character

Gana sorts the twenty-seven nakshatras into three groups of nine, and the group belongs to the star rather than to the person born under it. Of the eight kootas this is the one that reaches a family conversation already translated, and translated badly. Rakshasa gets rendered as demon, a classification of stars turns into a verdict about somebody’s nature, and a grandmother who would never say a cruel thing about a stranger repeats it because she believes she is quoting a text.

What does gana actually classify?

A nakshatra, and nothing else about the individual enters the calculation. The three groups are Deva, Manushya and Rakshasa — divine, human and the third, which we will come to. Nine stars in each, fixed, published, identical in every source.

The nine Rakshasa nakshatras are Krittika, Ashlesha, Magha, Chitra, Vishakha, Jyeshtha, Mula, Dhanishta and Shatabhisha. Read that list slowly, because it does most of the work of this post on its own.

How to read the reference below: each column is one neutral nakshatra category, with nine birth stars in each. The grouping belongs to the star; it is not a personality scale and it does not rank the people born under those stars.

Chitra and Vishakha are among the most common birth stars in India. Krittika is the star of Kartikeya. Magha is the star of the ancestors, and the fortnight associated with it is when a great many families perform their rites for the dead. These are not obscure positions and they are not treated as inauspicious anywhere else in the tradition. A person born in Magha is not held to be spiritually compromised by anyone; they are held to be born under the star of the pitrs. It is only inside this one koota that the label acquires the weight families give it.

Nothing in the assignment is derived from the individual. Not the ascendant, not a planet, not the birth time beyond the few hours that fix which nakshatra the Moon occupied. Two people born under Ashlesha share a gana whatever else is true of them: different decades, different countries, different everything. You can look up either one in the nakshatra finder without supplying a single fact about the person.

Gana is a property of a nakshatra. It becomes a property of a person only in translation.

Where did “Rakshasa” pick up the meaning it now carries?

In the Sanskrit corpus the three ganas are a familiar tripartite scheme, the same three-way division of temperament and mode that appears across the older texts, including in the nakshatra descriptions of the Brihat Parashara Hora Shastra and the character material in the Saravali. It is a category of being, in the way that deva and asura are categories of being in the epics, and it carries no implication that a human born under such a star belongs to that class of being. What reached the drawing room is an English gloss that flattened a classificatory term into a moral one.

How that happened is straightforward and hard to undo. Nineteenth and twentieth century English renderings needed a word, and demon was available. Then matching software needed a label to print in a report, and the English word was already there. A family reading “Rakshasa gana” on a printout in 2026 hears the word demon whether it appears on the page or not, and no amount of technically correct explanation gets fully underneath that.

The same drift ran through the word dosha itself, which once marked an imbalance to be weighed and now lands as a defect to be feared. That is set out in the history of the word dosha, and it explains why every finding in this system arrives sounding more final than the sources intend.

There is a smaller point here, and it is one worth being blunt about. If gana genuinely described a person’s character, you would expect the tradition to apply the label consistently — to say the same thing about a Magha-born person in a career reading as it does in a marriage reading. It does not. Outside matching, the gana of a birth star barely surfaces at all. A category that only becomes a judgement when a marriage is on the table is doing social work, not descriptive work.

How often does gana score zero, and why is the result not symmetric?

Full marks come up in 33.3% of all pairings and a zero in 22.2%. Three equal groups produce nine possible gana combinations, and the three matching ones — Deva with Deva, Manushya with Manushya, Rakshasa with Rakshasa — take all 6 points, which is where the full-marks rate comes from.

The zero is the interesting figure, and it is smaller than families expect. Only two of the nine combinations score nothing under the default matrix, and both require a Rakshasa bride paired with a groom of a different gana. Reverse the couple, Rakshasa groom with a Deva or Manushya bride, and the same table awards a point instead of none. That is a strange result for a rule supposedly describing compatibility between two temperaments, since the two temperaments have not changed.

Worth pausing on what the middle of the table looks like too, because families rarely see it. A Deva bride with a Manushya groom scores five out of six under the default matrix. Five, not zero. The overwhelming majority of mixed-gana pairings land in that band, a point or two down and nowhere near a refusal, and yet the phrase “gana does not match” gets used for all of them without distinction. A one-point deduction and a six-point one are described in the same sentence, and only one of the two can plausibly decide a marriage.

The asymmetry is not an artefact of the software. Raman states it plainly: a Manusha or a Deva man should not marry a Rakshasa girl, but marriage between a Rakshasa man and a Deva or Manusha girl is passable. Whatever one makes of that rule, it is what the standard table encodes, and its practical consequence is that the koota’s zeros are generated almost entirely by the bride’s birth star. The mean score across every pairing sits at 3.3 of 6, so the component is not usually where a total goes wrong. When it does go wrong, it goes wrong in one direction.

Five published tables, disagreeing where it counts

There is no single gana matrix. Five circulate, and they disagree precisely in the cells that produce zeros. This is the part of the subject nobody mentions in a consultation, and it is the part that should decide how much weight you give a gana finding.

Written with the bride’s gana as the row and the groom’s as the column, in the order Deva, Manushya, Rakshasa:

• The Astro-Vision table, which cites Muhurta Chintamani, carries Raman’s asymmetry: a Rakshasa bride scores nothing against a Deva or Manushya groom, while a Rakshasa groom against either scores one. Two zeros. This is the default here.
• The Saravali and Maitreya table encodes the opposite asymmetry. There it is the Rakshasa groom who zeros against a Deva or Manushya bride, and the Rakshasa bride against a Deva groom who scores one. Three zeros, in different places.
• The VedicMarga table is fully symmetric. It carries none of the asymmetry the others defend, which means it also answers the question of which partner is Rakshasa with a shrug.
• The FreeHoroscopesOnline table, attributed to the Charak lineage, and the AstroSight table each differ again — one cell apiece, and in each case a cell that turns a passing score into a zero or the reverse.

Consider what that means for a couple. Under one published table a pairing scores zero and the family is told the match carries gana dosha. Under another published table, with the same two birth stars and no change to anything else, the same pairing scores a point and the dosha does not exist.

Neither astrologer has made an error. They are reading different books.

A judgement about two people’s natures does not flip when you change which edition is on the desk. A table lookup does. That is the whole argument of this post, and the five matrices are the evidence for it — you can run all of them against your own stars in the gana tool and watch the verdict move.

A transposed table prints a plausible wrong answer

The published matrices are printed with rows as the bride and columns as the groom. That convention is not universally stated on the page: some sources give the table and leave the orientation to be inferred from surrounding prose, which is fine for a reader who knows the tradition and disastrous for a programmer transcribing it at speed. Transpose the table by accident and the symmetric cells stay put while every asymmetric one swaps. The total still looks entirely plausible, and nothing on the report indicates that anything went wrong.

The result is a report that is internally consistent, arithmetically correct given its own inputs, and wrong. A couple can be told they score zero on gana when the same table, read in the orientation it was printed in, awards them a point. There is no error message for this, because from the software’s point of view nothing failed. It is the sort of defect that survives testing indefinitely, because the only test that catches it is comparing against a printed source by hand.

Which is a reasonable argument for wanting to know which table an engine uses and in which orientation, not only what score it produced. The same class of transcription problem shows up in the yoni matrix, which is larger and therefore worse.

There is a cheap test for it, and anyone can run it. Take the two birth stars and swap which partner they belong to. If the score changes, the engine is applying an asymmetric table — fine, and it should say which. If the score changes in a direction that contradicts what the source you were quoted says about Rakshasa brides and Rakshasa grooms, something has been transposed. Two runs, thirty seconds, and a defect that no amount of reading the report would ever have revealed. It is not a clever test. It is just one nobody thinks to run, because the report arrives looking finished.

Why does South India grade gana instead of scoring it?

Because in the ten-porutham system gana is not six points out of thirty-six at all. It is a graded verdict with four levels, and the levels do not map onto the numbers. Deva with Deva is ati-uttamam, the highest grade. Manushya with Manushya is uttamam. A Rakshasa groom with a Deva or Manushya bride is madhyam, middling and permissible. A Rakshasa bride with a groom of another gana is adhamam, and rejected outright.

The case that repays attention is Rakshasa with Rakshasa. Technically it grades as uttamam, since the two ganas match and matching is what the grade rewards. South Indian practice nonetheless does not recommend it, and the recorded reason is a statement about how the marriage is expected to go rather than anything about the grade.

Hold those two facts together, because the gap between them is real and it is exactly what a numeric report destroys. The table passes the pairing. The practice declines it. A family in Tamil Nadu who consulted an astrologer heard the second thing. A family who ran the same two stars through matching software saw six points out of six and no comment at all. Both were told the truth about the tradition, and they were told different truths.

How the ten poruthams work as a system, and why they are not the eight kootas with two extra items bolted on, is set out in the porutham explainer.

Seven cancellations, one with a named authority

Seven documented rules are said to set gana aside. One carries an explicit statement from Raman and the rest travel on popular authority. The engine cites the text each rule comes from and reports which fired; whether a cancellation changes anything is not a claim made here.

• The bride’s star lies beyond the fourteenth from the groom’s. Raman is explicit that where this holds, the evil may be ignored. It is the only gana cancellation in the registry with a named modern authority stating it as permission rather than as a report of what others believe. You will find it in the 1948 edition of Muhurtha.
• A shared birth-star lord. The circulated verse grahaikye tu gaṇadveṣam covers this. It is the third clause of a couplet whose other clauses do the same job for nadi and bhakoot, it is quoted constantly, and it has not been located in a named printed edition of any of the texts it is attributed to. Reported here as traditional and flagged as unverified wherever it applies.
• Moon-sign lords the same or mutually friendly. Popular sources.
• Navamsha lagna lords the same or friendly. Popular sources, and one that needs a birth time precise enough to fix an ascendant — so for a great many charts it cannot be evaluated at all, which is not the same as failing.
• Same Moon sign but different birth stars. Popular sources.
• No bhakoot dosha present.
• The two nadis differ.

Those last two deserve a raised eyebrow. A rule that cancels gana because a different koota happened to pass is not an exception to gana at all: it treats the six points as a summary of findings made elsewhere and not as an independent measurement. If the component can be waved away whenever bhakoot came out clean, it was never carrying six points of its own information. That is a coherent position for a tradition to hold. It is just not the position anyone announces when they deliver the finding.

What does gana look like on the published sample couple?

Priya was born in Ashwini, a Deva star. Arjun was born in Purva Bhadrapada, a Manushya star. Different ganas, so the matching-pair cell does not apply, and the pair takes five of six points under the default table and all six under another published one.

Under the Astro-Vision table the Deva-bride, Manushya-groom cell holds five, so they lose a single point. Under the Saravali and Maitreya table the same cell holds six and they lose nothing. Under the remaining three tables it is five again. One point of movement, produced by nothing but the choice of printed source, on a couple whose total sits at seventeen of thirty-six — one point below the conventional accept line.

Follow that through. Their gana is not what refused them; the eight points nadi took did that, and their bhakoot came out clean. But the single point that moves between two published tables is the same size as the gap between their score and the line. On a different couple, with a different combination, that one point is the whole verdict. The full breakdown across all five traditions is in the guna milan tool, which shows the score, the table used and the rules that fired beside the number.

What can you check for yourself before accepting a gana refusal?

More than with any other koota — both birth stars are already printed on the report you were given, and the gana of each is a lookup in a published list of twenty-seven. That makes this the rare finding a family can verify at the kitchen table without an astrologer’s help.

• Which gana matrix are you using, and what does it cite? Five are in circulation and they produce different zeros.
• In that matrix, are rows the bride or the groom? A transposed table inverts every asymmetric cell and prints a perfectly plausible total.
• Is this being scored out of six or graded as a porutham? The two systems reach different conclusions about the same pair of stars, and one of them can pass a pairing the practice declines.
• Which cancellation are you applying, and is our nakshatra distance beyond fourteen? That is the one with a named authority behind it.

None of this asks anyone to abandon a tradition, and none of it requires arguing with an elder about whether astrology works. It asks a narrower question: given that five published tables disagree about this exact cell, which one produced the sentence that was said about my future wife, and does the person who said it know there are five?

The honest answer, most of the time, is that they do not — not through any dishonesty, but because software prints one number and the number does not carry its own provenance. Which is a fixable problem, and a much smaller one than the problem a family thinks it has been handed.