When was the abacus invented? Why sources say 2700 BCE, 2500 BCE and 300 BCE

The oldest physical counting board that survives is the Salamis tablet, a marble slab found on the Greek island of Salamis in 1846 and usually dated to about 300 BCE. The far earlier date you see in timelines — typically written "c. 2700–2300 BCE" — does not refer to a surviving abacus at all; it refers to the Sumerian and Old Akkadian period in Mesopotamia, when scholars infer from the sexagesimal (base-60) number system and administrative tablets that scribes were already reckoning with tokens or ruled columns. No Mesopotamian abacus has ever been excavated. Both dates can be quoted honestly, provided you say which question you are answering: earliest inferred use (3rd millennium BCE) or earliest surviving artefact (c. 300 BCE).

Where the "2700–2300 BCE" figure actually comes from

The range 2700–2300 BCE is not a measurement of an object. It is the conventional span of the Sumerian and Old Akkadian periods in southern Mesopotamia, and it is attached to the abacus because that is the era in which the sexagesimal (base-60) number system is attested in cuneiform administrative tablets. The argument, made by historians of number such as Karl Menninger and Georges Ifrah, is indirect: a scribe recording large quantities in successive base-60 orders is doing something that a ruled board with columns of counters performs naturally, and the layout of some accounting tablets mirrors columnar reckoning.

That is a reasonable inference, and it is why the date propagates through timelines and textbooks. But it is worth stating plainly what it is not. There is no excavated Sumerian or Babylonian abacus. No board, no set of counters identified with certainty as calculating counters, no contemporary text that unambiguously describes the instrument. When a timeline prints "abacus, c. 2700 BCE" with no qualifier, it is quoting a scholarly reconstruction as though it were an artefact date.

The neighbouring figure of c. 2500 BCE, which some sources use instead, is simply a rounder point inside the same window. It carries no separate evidence.

The Salamis tablet: the oldest counting board we can actually examine

The Salamis tablet was found on the island of Salamis, near Athens, in 1846, and is held in the Epigraphical Museum in Athens. It is a slab of white marble roughly a metre and a half long, incised with parallel lines and marked along its edges with Greek acrophonic numerals — the older system in which the sign for five, ten or a hundred is the first letter of the number word. Counters or pebbles laid in the ruled fields represented values; the Greek word for pebble, psephos, and its Latin equivalent calculus, are the ancestors of the verb "calculate".

Its date is usually given as about 300 BCE, sometimes stated more loosely as the 4th century BCE. That date is itself an estimate: the tablet was not recovered from a securely stratified excavation, so it is dated on the style of its numerals and its resemblance to other Hellenistic material. It is a good estimate, but it is not a laboratory date, and a careful page should say so.

Independent confirmation that Greeks used such boards comes from images rather than objects — most famously the Darius Vase, an Apulian red-figure krater of the later 4th century BCE, which shows a treasurer working at a table with counters and numeral signs. Herodotus, writing in the 5th century BCE, also compares Egyptian and Greek habits of reckoning with pebbles.

What came after: Roman, Chinese and Japanese forms

The abacus is not one invention with one date; it is a family of devices repeatedly reinvented in more portable form.

The Roman hand abacus is the first genuinely pocket-sized version: small bronze plates with beads sliding in slots, a handful of which survive from the imperial period, roughly the 1st century CE onwards. Its arrangement — several beads below a bar and one above — is functionally the ancestor of the East Asian designs.

The Chinese suanpan, with beads on rods in a frame, is securely documented from the medieval period; its unambiguous depictions and descriptions cluster in the Song dynasty and later, and it is fully systematised in Cheng Dawei's Suanfa Tongzong of 1592. Earlier claims — notably a passage attributed to Xu Yue in the 2nd century CE — are contested, because the surviving text is known only through much later commentary and the passage may describe a different reckoning method. The Japanese soroban derives from the suanpan, was in wide use by the 17th century, and was later simplified to one upper and four lower beads.

For anyone building a technology timeline that runs from the abacus to Jacquard's loom (1804–05) and on to Babbage, the honest framing is this: counting boards in Mesopotamia by the 3rd millennium BCE on inference, a surviving Greek board by c. 300 BCE, bead frames from Rome onwards, and programmable punched-card control only in the 19th century.

What most scholars accept, and what stays contested

Accepted: The Salamis tablet exists, was found on Salamis in 1846, is held in the Epigraphical Museum in Athens, and is the oldest surviving counting board known. Greek and Roman counting with pebbles and bead-slot devices is securely attested by both objects and texts. The Mesopotamian sexagesimal number system is attested in cuneiform tablets of the 3rd millennium BCE.
Mainstream: Most historians of mathematics accept that some form of counting board or token-based reckoning was in use in Mesopotamia well before the Greek examples, plausibly in the 3rd millennium BCE, because the base-60 system and the structure of administrative accounting presuppose a physical aid. The Salamis tablet is conventionally dated to about 300 BCE.
Contested: Whether a true abacus — a ruled board with counters in place-value columns — existed in Sumer or Babylonia at all is genuinely disputed, because the evidence is entirely inferential and no such object has been excavated. The precise date of the Salamis tablet is a stylistic estimate rather than a fixed date. The claim that the Chinese suanpan dates to the 2nd century CE rests on a text known only from later transmission and is not accepted by all specialists.

Frequently asked questions

Is there a real Babylonian abacus in a museum?

No. No Sumerian or Babylonian abacus has ever been excavated. The 2700–2300 BCE date attached to the abacus in timelines refers to a scholarly inference from the Mesopotamian base-60 number system and administrative cuneiform tablets, not to a surviving object. The oldest counting board that physically survives is the Greek Salamis tablet, dated to about 300 BCE and held in the Epigraphical Museum in Athens.

Why do some sources say 2700 BCE and others 2500 BCE?

Both figures point to the same window — the Sumerian and Old Akkadian period in Mesopotamia — and neither is derived from an artefact. 2700–2300 BCE is the fuller conventional range; 2500 BCE is simply a rounded midpoint inside it. Because the date describes an inferred practice rather than a dated object, no single year can be defended, and a range is the accurate way to state it.

How old is the Salamis tablet and how is it dated?

The Salamis tablet is usually dated to about 300 BCE, sometimes given more loosely as the 4th century BCE. It was found on the island of Salamis in 1846 and is now in the Epigraphical Museum in Athens. Its date is an estimate based on the style of the Greek acrophonic numerals carved on it, not on excavation context or scientific dating, so the figure should be quoted as approximate.

Explore related content on this site

Sources and further reading

Encyclopaedia Britannica's entry on the abacus, covering the Salamis tablet, the Roman hand abacus, and the East Asian suanpan and soroban.
The standard scholarly account of counting boards, pebble reckoning and the Salamis tablet, with detailed discussion of Greek and Roman practice.
A wide-ranging survey of reckoning instruments that sets out the inferential case for Mesopotamian counting boards alongside the surviving Greek and Roman evidence.
Museum collections of abaci and counting devices, including Chinese suanpan and Japanese soroban examples, useful for the later history of the instrument.
The Epigraphical Museum in Athens holds the Salamis tablet itself; its catalogue is the primary reference for the object's provenance and dating.