Have you heard about the two Israelis who argued whether this week’s fast day of the Ninth of Av should be called Tisha b’Av or Teysha b’Av? In the end they compromised on August 13.

Of course you haven’t, because I just made it up. Yet such an imaginary anecdote would illustrate something real about Hebrew. This is a feature that confuses many of its native speakers, let alone those who study or acquire it as adults—namely, that while it is a heavily gendered language in which all nouns and their agreeing adjectives (as well as most verbs and some adverbs) are either masculine or feminine, Hebrew numbers come reversed, so that masculine nouns take feminine numbers and feminine nouns take masculine ones. Numerically speaking, it’s an upside-down world.

Let’s be more concrete. The Hebrew suffix –ah (in which the “h” is unpronounced but indicates a stressed syllable) is what most commonly converts a masculine noun or adjective into a feminine one. “Boy,” for example, is yeled, “girl,” yaldah; a big boy is yeled gadol, a big girl, yaldah g’dolah. Logically, therefore, since the masculine form of the Hebrew number “nine” is teysha and the feminine form is tish’ah, and since the Hebrew word for “day,” yom, with its plural of yamim, is masculine, “nine days” should be teysha yamim. (The suffix -im is the regular masculine pluralizer in Hebrew, just as the suffix -ot is the regular feminine pluralizer.)  And yet, as we have said, it’s just the opposite: “nine days” is tish’ah yamim, whereas “nine minutes” is teysha dakot, even though dakah, “minute,” is a feminine noun.

Go explain it! Many Israelis who can’t simply don’t bother with the distinction between masculine and feminine numbers and use the feminine ones, with their masculine appearance, for everything. If it’s nine of something, then it’s teysha in every case: teysha dakot and teysha yamim, teysha y’ladot (“girls”) and teysha y’ladim (“boys”), even though, in terms of formal Hebrew grammar, teysha yamim and teysha y’ladim are incorrect. These same Israelis will also say shalosh sh’kalim, “three shekels,” instead of sh’loshah sh’kalim; shesh tapuḥim, “six apples,” instead of shishah tapuḥim, and arba pilim, “four elephants,” instead of arba’ah pilim. The purists cluck their tongues at this? Let them cluck!

Indeed, the division between those who “correctly” say sh’loshah sh’kalim and those who “incorrectly” say shalosh sh’kalim forms a fault line in Israeli society between a more educated class of speakers who try to observe Hebrew’s traditional grammatical rules and a less educated class that couldn’t care less about them. Say “three shekels” in Hebrew and you’ve probably told me a lot about the schooling you’ve had, your socio-economic status, the kind of neighborhood you live in, perhaps even your politics. I say “probably,” because there’s no lack of exceptions. There are uneducated Israelis who wish to speak educatedly, educated ones who like to sound folksy, and many who shift back and forth. The same Israeli who says sh’loshah sh’kalim in one situation may say shalosh sh’kalim in another; he may use the former with the students he teaches at the university and the latter with the mechanic who fixes his car, or correct his twelve-year-old daughter when she says shalosh rather than sh’loshah but not his five-year-old son. There’s a lot of inconsistency.

There’s almost none, however, when it comes to the more general issue of gender in Hebrew nouns, adjectives, and their plurals, even though these, too, can seem maddeningly inconsistent to the non-native. Of course, most languages with gendered nouns and adjectives demand a significant amount of rote learning. If you’re studying French, you have to remember that soleil, sun, is masculine and étoile, star, is feminine, and that you say un soleil brillant, “a bright sun,” but une étoile brillante, “a bright star”—and you have to do this with thousands of words. But you don’t have to worry about plurals, because both masculine and feminine plurals are formed by the same suffixed “s.” It’s les soleils brillants, “the bright suns,” and les étoiles brillantes, “the bright stars,” and all the thousands of other cases are the same.

Not so in Hebrew. Here, quite apart from the fact that many words when pluralized can change some of their vowels and consonants, or even take infixes rather than suffixes, there is not only a masculine plural ending and a feminine plural ending; there are also a very large number of masculine nouns that are pluralized femininely and a smaller but still considerable number of feminine nouns that are pluralized masculinely. It’s bewildering.

Take, for instance, the two similarly structured words tsipor, bird, and tsinor, pipe or hose. There’s nothing about them to tell you that tsipor is feminine while tsinor is masculine, so that one says tsipor g’dolah, “a large bird,” but tsinor gadol, “a large pipe”; yet feminine tsipor pluralizes as tsiporim, with the masculine -im suffix, while masculine tsinor pluralizes as tsinorot, with the feminine -ot suffix! And just to add to the confusion, their accompanying adjectives do not follow them in this, so that “large birds” is tsiporim g’dolot and “large pipes” is tsinorot g’dolim. This kind of thing is quite normal in Hebrew and can reach seemingly irrational extremes, as when a word like av, father, pluralizes femininely as avot and a word like ishah, woman, pluralizes masculinely as nashim. The fact that Israelis are unfazed by such things and almost never get them wrong once past childhood is a tribute less to them specifically than to the overall human capacity for learning languages, which can, when the learner is young enough, cope effortlessly with irregularities of all kinds.

But what is the explanation of Hebrew’s wildly unpredictable gender system? What developments can account for it? A great deal has been written on the subject over the years by historical linguists and Semiticists, for the problem is one of Semitic languages in general and not of Hebrew alone. (Colloquial Arabic, for example, has the same gender reversal in its numbers that Hebrew does, and has gone even further in the tendency to use the feminine form of the number exclusively.) Although the scholars don’t agree about everything, most by now concur that this involved at least five stages in proto-Semitic, the common ancestor, spoken thousands of years before the Common Era, of Hebrew, Phoenician, Aramaic, Arabic, Akkadian, Ge’ez (the parent of Amharic and other Ethiopian languages), and still additional Semitic tongues.

In Stage 1, conjecturally, there were few or no pluralized nouns. (There were, however, masculine and feminine adjectives, the latter marked by the suffix -t or -at.) Evidence pointing to the absence of plurals can be found in biblical, mishnaic, and even modern Hebrew, in which enumerated groups of eleven or more units often take a singular noun. Thus, when we read in Genesis 5:27, “And all the days of Methusaleh were nine-hundred-and-sixty-nine years,” the Hebrew literally says “nine-hundred-and-sixty-nine year,” with the word for year, shanah, in the singular. In the distant proto-Semitic past of which such a usage is a remnant, it is presumed, a word like shanah had no plural form at all.

In Stage 2, the proto-Semitic numerals three through ten (one and two had a different history) developed a collective form marked by the same -t or -at suffix and having the sense of “a group of.” Thus, the number eight with such a suffix came to mean “a group of eight” (of whatever noun followed), and so on. Eventually, for unclear reasons, these collectives with their -t or -at suffixes became ordinary masculine numbers, whereas the feminine numbers remained unsuffixed.

In Stage 3, the collectivizing -t or -at suffix became a general plural marker and all or most nouns, both feminine and masculine, were pluralized by means of it.

In Stage 4, a masculine plural suffix, -im or -in, began to develop in some of the Semitic languages, while -at or -ot became a distinctly feminine ending. In Hebrew, most masculine nouns that were previously pluralized by -ot switched to -im, but the process failed to affect them all. Hence, Hebrew’s many masculine nouns with feminine plural endings, like av/avot.

In Stage 5, which belonged to Hebrew alone, the feminizing -at suffix was aspirated into -ah. (In what are called construct forms, however, the -t was retained. Hence, yaldah, a girl, but yaldat k’far, a “girl of the village” or country girl.) This aspiration also affected the -at ending of masculine numbers, which now also ended with -ah. Their feminine appearance is thus the ultimate result of the processes that took place in Stage 2, as is the masculine appearance of the feminine numbers.

In Stage 6, . . . ah, but we’re still in Stage 6 today! Will there be a Stage 7 in which the masculine numbers will have disappeared from Hebrew entirely? We should know the answer another 50 or 100 Tisha b’Av’s from now.