Before any mark was placed above or below a Hebrew letter to guide the voice, the melody already existed. It lived in the throats of the readers, learned by listening, carried in the body the way a person carries a mother tongue. When Yehuda HaNasi gathered the sages at Beit She'arim around 200 CE to fix the text of the Mishnah, the Torah readings in the synagogues of the Galilee and Judaea were already ancient performances - chanted without notation, from memory, in a way that had not changed in living memory or in the memory before that.
This is the era of chanting that is not yet written down. The Masoretes who would one day encode the melody into a system of marks - the te'amim, the trope - were still five centuries away. The codices they would annotate had not yet been made. What existed instead was something more intimate and more fragile: a tradition of sound, alive only so long as a teacher passed it to a student and the student did not forget.
The record
The Babylonian Talmud (Megillah 32a, redacted c. 500 CE but preserving earlier rulings) states that one who reads the Torah without a melody - ha-kore b'lo nima - shows disrespect for it, and one who expounds it without a tune commits a similar failing. The passage cites Ezekiel 33:32, "a lovely song with a sweet voice," as proof that musical delivery is an obligation, not an ornament. The ruling presupposes that a distinctive chant already existed and was expected - long before any marks existed to prescribe it. See Talmud Bavli, Megillah 32a.
The question scholars cannot fully answer is how consistent the melody was. Modern cantillation traditions - Yemenite, Ashkenazi, Sephardi, Iraqi - share the same written marks yet sing them in melodies that sound quite different from one another. Those differences almost certainly preserve something real: distinct local chant styles that existed before the Masoretes, which the marks then fixed in place without fully unifying. The written system, when it came, crystallised what was already a branching oral tradition. The root went back to the synagogue chant of the Mishnaic period; the branches had already diverged.
The record
The earliest explicit description of how to transmit Torah chant appears in the Masoretic treatise Diqduqe ha-Teamim, attributed to Aaron ben Moses ben Asher of Tiberias (fl. c. 900-950 CE). Ben Asher's opening lines make clear that the system he is codifying is not new: he presents himself as fixing what has been received, not inventing it. The treatise is preserved in the same manuscript tradition as the Aleppo Codex and has been edited critically by Aron Dotan (1967, Hebrew University of Jerusalem). The gap between ben Asher's written system and the Mishnaic-era oral chant it encodes is roughly 700 years - seven centuries in which the melody lived entirely in the human voice.
There is something striking about the fact that the tradition survived those centuries intact enough to be worth encoding. No music existed on parchment, yet the communities knew the tune. What held it was the weekly reading itself - the Torah portion chanted every Shabbat, year after year, in a rhythm that made the melody as ordinary and as essential as breathing. A child heard it before she could read. By the time she stood at the reading desk herself, the chant was already part of her. The marks, when they came, only confirmed what the voice already knew.
The record
The musicologist Abraham Zvi Idelsohn argued in his Jewish Music in its Historical Development (1929, Holt, New York) that certain melodic patterns in synagogue chant are demonstrably ancient - pre-Masoretic in character, paralleled in the liturgical music of communities that separated from the main body of world Jewry before the Masoretic marks were standardised. His comparisons of Yemenite, Persian, and Ashkenazi cantillation - based on field recordings he made in Jerusalem between 1911 and 1913, now held in the National Library of Israel - pointed to a shared melodic core that must predate the division of those communities. The conclusion is that what we hear in the synagogue today carries, in its bones, something from the chanting of the Mishnaic period.
The marks came later. The melody came first. And before either of them, there was a people who had decided, for reasons they may not have been able to articulate, that sacred words should be sung rather than merely spoken - that the voice was part of the text, not an addition to it. By the time the Masoretes set pen to parchment, they were not writing music. They were writing down something they had already been singing for longer than anyone could remember.
