Professor Aaron Koller of Yeshiva University (and soon to be in Cambridge) has written a masterful article on this topic and I am going to share some insights from it. He analyzes when markings between words (and later, spaces between words) arose in Hebrew and other Semitic languages. A main point he makes is that where one word ends and another begins has always been a thorny problem. Thus, separating between “words” is not a simple issue. The article is entitled: “The Evolving Definition of ‘Word’ in Early Northwest Semitic Writing: From דברים ot תיבות,” in the Journal of Near Eastern Studies 83:1 (2024). In English, for example, he asks, why is “birthday” one word, but “wedding day” two separate words.

Koller is primarily interested in word division in alphabetic languages like Hebrew and Aramaic, but he first writes about the practice in cuneiform languages and Egyptian hieroglyphs. On those languages (already in use in 3000 BCE), he concludes: “The two major writing systems in the Near East diverged when it came to the question of word dividers. Cuneiform did not systematically visually mark the divisions between words, making rapid silent reading difficult … In Egypt, on the other hand, word division was built into the writing system from the beginning, taking the form of obligatory classifiers in hieroglyphic, and a more pronounced form in hieratic. This must have enabled scribes to identify words quickly and read even unfamiliar texts with fluency.”

The present scholarly view is that alphabetic writing arose in the early second millennium BCE, and that it seems to have originated in the area of the Sinai and of Egypt. (The shapes of the letters were influenced by the shapes of Egyptian hieroglyphs.) The earliest alphabetic writings that we know of did not have word dividers.

For example, recently discovered was a comb from Lachish—made of ivory—from around 1700 BCE. (I wrote about this before. The comb had the following language on it: “May this tusk remove the lice of hair and beard.”) This comb did not have word dividers. Some of these ancient writings without word dividers were a type of graffiti. Others were not meant for human reading. For example, a dedicatory inscription to the goddess Hathor. The primary audience for this writing was the goddess herself. Koller writes: “While we have no specific information about the goddess’s level of literacy, it is reasonable to assume that she could read even without spacing.” But eventually, as alphabetic writing became meant for other humans to read, systems for marking word divisions arose.

Already in the 16th century BCE, on the Tell Nagila shard (in southern Israel), we have a vertical sign between two words. On a bowl from Lachish in the 13th century BCE, there is a similar sign. On a jug from Lachish from the same period, we have three dots used as a division between words. A 12th century BCE bowl from Qubur al-Walaydah (southwestern Israel) has lines between words.

Ugaritic—spoken in ancient Syria—was a language written in cuneiform, but (unlike Eastern Semitic cuneiform) it was an alphabetic cuneiform. Each cuneiform diagram represented a letter (the same letters as Hebrew, but a few more). Thus, it is relevant too. In Ugaritic texts from around the 13th century BCE, a vertical wedge was used to divide words.

Koller summarizes: “Thus, most Late Bronze and Iron Age alphabetic texts show some marker of word division: three dots, one dot, a wedge or a line. While it is easy to observe the introduction of word dividers, it is much more difficult to discern systematic rules governing their use.” For example, he cites one scholar who believes three different systems were in use in Ugaritic.

Koller gives many examples of why defining a word is a thorny issue. I will discuss only two:

In English, we have words like “in,” “to” and “and.” These are written as separate words. In Hebrew, those meanings are found in the single letters “bet,” “lamed” and “vav.” An issue is whether those Hebrew letters—when used with these meanings—should be written as separate words. Eventually, the practice developed not to write as separate words those words that would only be one letter. But it did not have to evolve this way.

What about לבבכם? It means “your heart.” In English, this concept is described with two words. What position should Hebrew take on this issue? In Aramaic and Moabite, the ending is הם, and הם was, sometimes, written as a separate word. But in Hebrew, we never have כם as a separate word.

Koller summarizes: “The scribal practice of dividing words is not in fact as old as the alphabet, but instead developed erratically over the following 1,500 years … The reason word dividers were not born with the alphabet is that the idea of word division is not as intuitive as is sometimes thought … Orthographic practices that developed in the Iron Age reflect a negotiation between linguistic intuitions and scribal practices.”

Koller concludes: “All the Imperial Aramaic texts utilize spacing to separate words. Jewish scribes too, adopted the Aramaic script and scribal practices during the Persian era. Thus, systematic spacing became the norm in Hebrew and Aramaic texts by the middle of the first millennium BCE.” (By “Imperial Aramaic,” he means Aramaic starting from the 7th century BCE. We have evidence for spacing from the Aramaic of Mesopotamia in this century.) For example, there are consistently spaces between words in the Aramaic papyri from Elephantine (in Egypt) from the fifth century BCE.

Another scholar summarizes what happened in the northwest Semitic languages as follows: “Initially, words were divided with a simple vertical line, which, over time, shortened and evolved until it became a simple dot … At Ugarit, the word divider consisted of a small vertical wedge, and the word divider of the Lachish Ewer consists of three vertical dots. However, continuous writing persisted in Phoenician.” Why the Phoenician scribes continued to make things difficult for their readers is a mystery.

———

A few miscellaneous points:

What about the Hebrew word for “word?” In rabbinic Hebrew, we can find the word דבר (davar) used with the sense of “single word.” (Since a word is something that would be “spoken.”) See, e.g., Sifrei Deuteronomy 306, which states that Moses used the name of God in Deuteronomy 32:3 after 21 “דבר ’’ of the song. But another term arose, תיבה, which is the usual word for “word” in Mishnaic Hebrew. It was originally the term for a “box.” Probably, it is used because it reflects the idea that there are spaces all around (boxing in) the written word. Interestingly, this word for “word” originated in Egyptian! See B. Noonan, “Non-Semitic Loanwords in the Hebrew Bible,” page 217.

Within Jewish scribal circles there remained an association between the type of script and word division. Texts written in Old Hebrew script among the Dead Sea scrolls are almost always divided by dots, while texts in the square Hebrew script among the scrolls are divided by spaces. There is only one Old Hebrew text among the scrolls with division by spacing. Also, Old Hebrew texts sometimes split words between lines, while square script texts never do so.

Koller also writes: “In classical writing, spacing was introduced only in the early Middle Ages, beginning in Ireland.”

Koller also discusses the issue of whether the Torah was originally written in a continuous script without word divisions. The scholarly consensus now is that it was not.

In the context of writings that were meant to remain private, Koller mentions a story involving Barack Obama. As a senator in 2008, he left a note in the Kotel and then someone removed it and sold it to an Israeli newspaper! The rabbi of the Kotel sharply condemned the theft and said that it “represents a desecration of the holy site.”

———

I will end with a story about my daughter. After her high school years at Bruriah, she was reading many works in Hebrew by Acharonim about religious thought. She came across the word ותו. Having had little exposure to Tosafos (which use this word all the time to preface their second answer), she said to me: “תו cannot be a word. It is missing a root letter at the end.” I looked in Jastrow and, lo and behold, she was correct. It is a shortened form of תוב and is the equivalent of שוב (again, furthermore).

MitchellFirstcanbereachedatMFirstAtty@aol.com. He would like to thank Rabbi Reuven Chaim Klein for sending him the above article by Professor Koller.