Parashat Shoftim, and the numbers in it

Judges, witnesses, cities and a heifer, and a name whose usual spelling the portion never uses.

This coming Shabbat, the fifteenth of August 2026, is the second of Elul, and the reading is Shoftim: Devarim 16:18 through 21:9. It is a portion about courts, kings, prophets, witnesses and war, and it is one of the few in the Torah where the numbers are not decoration. Judges are counted. Witnesses are counted. Cities are counted. The Torah itself raises an arithmetical question in the middle of it and leaves the answer to the Oral Law.

Since this site would rather count than remember, here is the portion as the index has it: 97 verses, 1,523 words, 5,592 letters, and a standard value for every letter in it of 364,583. That last figure is the kind of thing only an index can tell you, and it is worth saying at once that it means nothing in particular. It is a measurement, not a teaching.

The name, and the spelling nobody uses

Start with the word the portion is named for, because it is a good example of how easily a gematria goes wrong before the arithmetic even begins.

Written out the way the name is usually written, with a vav, שופטים comes to שופטים=445. That is the number you will find if you look the parsha up almost anywhere.

It is not the number of the word in the verse. The Torah spells it defectively, without the vav, and that spelling comes to שפטים=439 instead. Six apart, and the six is the vav that is not there.

Our index is blunt about which of the two the text actually contains. The defective spelling stands 22 times in the twenty-four books. The full spelling, the one the name is printed with, stands 0 times. Not rarely. Never.

This matters beyond pedantry, because gematria is computed on the letters that are written. A value taken from a modern spelling of a biblical word is a value of a word the Bible does not contain, and every so often somebody builds an argument on one.

A second warning is folded into the same count, because not all of those occurrences are courts. The majority of them are not judges at all: the same four consonants carry a second word entirely, the plural of shefet, a judgment carried out, which is what Shemot and Yechezkel mean when they use it. The index counts spellings, because spellings are what gematria weighs. The reader who opens the door below will see both words standing in one list, in canonical order, and can sort them out by eye in about a minute.

Justice, twice

Two verses later comes the most quoted line in the portion, and the doubling in it is not a figure of speech that the text uses elsewhere.

צדק stands 70 times in the twenty-four books. This is the only verse in any of them where it stands beside itself. We looked, which is the whole point of having the index: the doubling that every commentator writes about is, as a matter of the masorah rather than of homiletics, unique to this line.

צדק=194

Rashi’s comment on the doubling is one short line in the Hebrew and it is not a gematria. Go after a fine court, he says, following the Sifrei: the second צדק is not emphasis, it is an instruction to the litigant. Pursue justice by choosing the beit din that will do it properly, even at the cost of a journey. A reading that turns a rhetorical repetition into a practical obligation is worth more than any number the word carries, and it is the kind of thing to keep in mind before reaching for arithmetic.

Two witnesses, or three

Here is the portion doing arithmetic on itself.

If two witnesses establish a capital case, the third is not needed, and the Torah does not usually name a number it does not need. The Mishnah in Makkot asks exactly that question, in those words: if the testimony stands with two, why did Scripture specify three? Its answer is that the verse is not counting witnesses at all. It is equating them. A group of three may convict a group of two of false testimony, and, which is the Mishnah’s point, a group of two may equally convict a group of three. The number in the verse is a rule about sets, not a tally.

Chapter 19 states the same principle in the general case, and the same Mishnah carries it out to any number of witnesses at all. Rabbi Akiva reads the third differently again: he came only to be stricter with him, to make an accessory’s judgment the same as a principal’s. Three readings, one number, and none of them arrived at by adding anything up.

Three, twenty-three, and why a court is never even

The mandate at the head of the portion, judges in all your gates, is where the numbers of the Jewish court system are hung. The Mishnah opens Sanhedrin with them. Monetary cases are heard by three. Capital cases are heard by twenty-three, and so, the same Mishnah adds without blinking, is an ox that gored, and a wolf, and a lion, and a snake.

The reason a court has an odd number is stated most plainly in a Mishnah about the very last law in this portion, the heifer whose neck is broken when a body is found between towns. Three elders come out from the high court in Jerusalem, says the first opinion. Rabbi Yehudah says five, and derives it from the two plural nouns in the verse, two elders and two judges, and then adds one more for a reason that has nothing to do with the verse: a court may not be evenly balanced. Four judges can deadlock. Five cannot.

And one number in Makkot that is really an argument. A Sanhedrin that puts a man to death once in seven years is called destructive. Rabbi Elazar ben Azariah says once in seventy years. Rabbi Tarfon and Rabbi Akiva say that had they sat on it, no one would ever have been executed, and Rabban Shimon ben Gamliel answers that they too would have multiplied bloodshed in Israel. The whole dispute is conducted in numbers and not one of them is a gematria. It is worth noticing how often that is the case in the parsha the numbers people like best.

Whole, not perfect

תמים is the word, and the JPS rendering above says whole-hearted rather than perfect, which is the better English by some distance. It stands 49 times in Tanakh, most often of an animal without a blemish and of a man who walks straight, and it comes to תמים=490.

Its position in the portion is the argument. The verse before it forbids the diviner, the soothsayer and the one who enquires of the dead; the verse after it promises a prophet. Between the two stands the demand to be whole, which Rashi glosses, following the Sifrei, as walking with Him wholeheartedly and not investigating the future, taking whatever comes and accepting it. The instruction sits precisely where a person would most want a forecast. That placement is the teaching, and no number is doing the work.

Three cities, and the hands that were washed

Two more counted things, and then the portion is done.

Three cities of refuge, set apart in the middle of the land, with roads prepared to them and the territory divided into thirds so that no fleeing man has too far to run. Three more are promised if the borders ever widen. The Torah is not being poetic with these numbers. It is describing a road network.

And the last law of the portion, which is also its strangest. A body is found in a field, the killer unknown, and the elders of the nearest town break a heifer’s neck in an untilled valley, wash their hands over it and make a declaration.

Both Rashi and the Mishnah in Sotah stop on the obvious objection. Would it enter anyone’s mind that the elders of the court are murderers? The declaration is not about the killing. It means: he did not come into our hands and get sent away without food, and we did not see him and let him go without an escort. The elders are clearing themselves of the indifference that made the road unsafe, not of the blow. A portion that opened by appointing judges in every gate closes by making the judges answer for a road between two towns.

On the remazim at the head of the page

Open a Chumash at this portion and the outer column carries the Baal HaTurim, as it does at the head of all fifty-four. It would be easy to reprint two or three of his comments here and easier still to invent one that sounds like him. Neither of his Torah commentaries is in this site’s library, so this article quotes none of them and gives no chapter or verse of them, for the same reason the article about him quotes none.

What is above is his method rather than his words. The three tools are gematria, notarikon and masorah, and the third is the one this portion rewards. That צדק stands beside itself in exactly one verse of the twenty-four books, that the name of the parsha is spelled in a way the parsha never spells it, that a word for judges is also a word for judgments: those are masoretic facts, they are checkable in a second on this site, and they were the work of a lifetime’s memory when he made them.

The Shabbat this falls on

Shoftim is read as Elul opens, and the coincidence is old enough that generations of darshanim have read the gates of the first verse inward, as the gates of a person, in the month when a person is meant to be looking at himself. That reading is not in the text and it is not in any book this site holds, so it is offered here as what it is, a homiletic commonplace rather than a source.

The plain calendar fact is firmer and more interesting. This portion, the one that appoints judges and courts and demands justice twice, is the one the year hands you on the first Shabbat of the month before the Day of Judgment. Nobody arranged that. The cycle simply lands there, and it lands there most years. What Elul does with the number is the next article.

The portion’s full text, its aliyot and its haftarah are at Shoftim in the reader, and every value of every word in it is at the portion’s gematria desk.