Scripture is full of numbers that clearly carry weight. Gematria is often reached for to explain them — sometimes rightly, more often not. It is worth separating the numbers that are simply symbolic from the ones where letter arithmetic is actually part of the text.
Seven: completeness
Seven is the most frequent significant number in the Hebrew Bible: seven days of creation, seven-year cycles, seven-branched lampstand, and in Revelation seven churches, seals, trumpets and bowls.
No gematria is required. Seven functions as a number of completeness in the surrounding cultures too, and the text uses it that way directly. Reading it as a letter sum adds nothing the passage was not already saying.
In English Ordinal, incidentally, the word seven comes to 65 — a fact about English spelling, and about nothing else. It is a good example of a calculation that is correct and irrelevant.
Forty: a full span of time
Forty days of rain, forty years in the wilderness, forty days of fasting. The pattern is consistent enough that "forty" reads less as a count than as an idiom for a complete period — long enough for something to be finished.
Again there is no letter arithmetic underneath it. The number is doing rhetorical work, not numerical work.
One hundred fifty-three: the interesting case
John 21:11 gives the exact count of fish in the net: 153. The precision is unusual, and readers have been trying to account for it since antiquity.
The mathematically real properties are genuinely elegant. It is the sum of the integers from 1 to 17 — a triangular number. It is also the sum of the cubes of its own digits, 1³ + 5³ + 3³ = 153. Augustine built an interpretation on the 1-to-17 sum; others have offered Greek isopsephy readings of phrases that add to 153.
This is where the honest boundary sits. The number is in the text. The arithmetic properties are facts. Which reading the author intended, if any, is not something arithmetic can settle — and the number of phrases that can be made to equal 153 is very large, which is exactly why a proposed reading needs external support to be worth much.
Six hundred sixty-six: gematria in the text itself
Revelation 13:18 is the one place where the text openly instructs the reader to count: it calls the number of the beast the number of a man, and gives it as 666 — with an old and well-attested variant reading of 616.
This is gematria proper, and the practice would have been familiar to a first-century audience: both Hebrew and Greek letters double as numerals, so counting a name was a normal thing to do. The most widely held scholarly reading takes the name back through Hebrew letters to Nero Caesar, whose Hebrew spelling sums to 666 — and whose Latin-derived spelling, one letter shorter, sums to 616. That a single dropped letter accounts for the textual variant is the strongest argument in the reading's favour.
Note what makes this case different from the others: the text asks for a calculation, the letters used are ones that carried numeric values for their own readers, and a second independent detail — the variant — falls out of the same explanation. That is a much higher bar than "these two English words add up the same."
Where English calculators fit
The word beast is 47 in English Ordinal, and bible is 30. Both are true and neither has anything to do with scripture, because neither Hebrew nor Greek letters were involved. English gematria is a modern practice applied to a modern language; it can be interesting on its own terms, but it cannot reach back into a Hebrew or Greek text.
If you want to work with the originals you need the letter values themselves — the Jewish gematria post lays out the Hebrew tiers and explains why the cipher labelled "Jewish" in English calculators is a Latin-alphabet imitation rather than the Hebrew system.
A working rule
Before reading gematria into a passage, ask three things:
- Does the text ask you to count? Revelation does. Genesis does not.
- Are the letters ones that carried numeric values for the original readers? Hebrew and Greek, yes. English, never.
- Does the reading explain a second thing you were not trying to explain? The 666/616 variant is the model.
Most proposed readings fail the first question. The few that pass all three are the ones scholars still argue about — which is a good sign that the questions are the right ones.