Semitic writing systems, such as those used for writing Hebrew, Aramaic, Ugaritic and Phoenician, are well known for the fact that signs for vowels are routinely left out. Have a look at the first line of the first book of the Bible, Genesis 1.1 (text taken from with the vowels and cantillation signs removed), the so-called ‘consonantal’ text:

בראשית ברא אלהים את השמים ואת הארץ

This is how this verse would have appeared in antiquity. The vowel points and cantillation marks that we find in Hebrew Bibles today came in in the medieval period (

בְּרֵאשִׁ֖ית בָּרָ֣א אֱלֹהִ֑ים אֵ֥ת הַשָּׁמַ֖יִם וְאֵ֥ת הָאָֽרֶץ׃

As the following transcription of the consonantal text shows, most of the letters correspond to consonants, and the vowels are largely unwritten (the main exception being the /ī/ vowel in rˀšyt = /rēšīt/ “beginning”):

brˀšyt brˀ ˀt ˀlhym hšmym wˀt hˀrṣ

Greek is famous for having taken a Northwest Semitic alphabet and introduced regular vowel writing (see for example Sampson 2015, 104–105). There is some evidence for believing that Greek may not have been the first writing system to introduce regular vowel writing—this honour may belong to Phrygian (see Waal 2020, 114). At any rate, at least from a typological point of view, it is clear that Greek (and Phrygian) writing differs from Northwest Semitic in that if a vowel appears in the spoken language, you have to write it down; in Northwest Semitic, you don’t have to.

Such a lack of interest in recording even the presence of a vowel might suggest a lack of interest in how the language sounds, in favour of recording the meaning. Is it the case, though, that Northwest Semitic writing systems are uninterested in recording ‘phonetic’ information? It is this question that I try to address in my recent monograph for the CREWS project, The Semantics of word division in Northwest Semitic writing systems: Ugaritic, Phoenician, Hebrew, Moabite and Greek.

There is a component of these writing systems – often overlooked – which turns out to pertain just as much to the language-as-spoken as the omission/inclusion of vowels. The component in question is the word divider.

We are accustomed to separating words by means of spaces. In the ancient Mediterranean world, the equivalent of the space was often the dot <·>. It may also appear as two dots placed on top of one another <:> or three dots <⋮>.


So much for the way of separating words on the page (or, as the case may be, stone or clay tablet), but what is a word anyway? It turns out that this question is harder to answer than it might first appear, and furthermore, that the ancients had a rather different view than we do.

When we put spaces between words, we are separating ‘words’ as we might look them up in a dictionary. The composers of ancient Greek, Hebrew, Phoenician or Ugaritic inscriptions (most of the time) didn’t separate words in this way. To see the kind of things we get, have a look at this inscription (from

Funerary inscription from Sicily (first or second century AD). Photo by M. Metcalfe, 2016-12-11, courtesy Museo regionale interdisciplinare di Messina

This is a funerary inscription from Sicily from the first or second century ad in Sicily (ISic001231; you can find the full record here: (As it happens, I don’t discuss this example in the book, but in an article that will shortly be coming out in another CREWS volume, Crellin forthcoming. In the book, there are lots of other examples from Ugaritic, Phoenician, Moabite and Hebrew in the book.) Here is a transcription (text per Crellin forthcoming.):

1   · Θ · · Κ ·






And a translation:

“To the underworld deities, Androbios Lukios, shipowner, lived without offence for 36 years. Erected by his brother Apollonios, with Mousaios and Theodoros, for the memory of our brother.” (trans. per ISic001231)

For the most part, the dots correspond to where we would divide the words. In a couple of places, however, we do not find word divisions where we would expect them. The most obvious is in line 5, where we have ΚΑΙΘΕΟΔΩΡΩ instead of ΚΑΙ·ΘΕΟΔΩΡΩ “and Theodoros”. The other in line 4, where we have ΣΥΝ “with” without a following dot, before ΜΟΥΣΑΙΩ “Mousaios” on the next line. (Even though there is a line break here, we would expect to find a word divider, since we have a word divider at the end of a line in line 3; conversely there is no word divider at the end of line 2, where the word carries over to the next line.)

Strikingly, you get exactly the same thing in written Hebrew. In the example from Genesis 1, there are three small words that are written together with the following words: ב b “in”, ה h “the” and ו w “and”.

So what is going on? Were the writers simply a bit careless? This has often been the view in the scholarship, especially in regard to ancient inscriptions. Either that, or that the writers were simply a bit confused about what a word is. But could there be a method in their (apparent) madness?

In both the Hebrew and the Greek cases, it is small, one-syllable words that lack word division. This gives us a clue as to what is going on. Let’s think about a sentence like:

I ate a cake for lunch.

In an ancient Greek (or Phoenician, Hebrew, or Ugaritic) inscription you get the equivalent of word division like this:


Why separate the words like this? Well, let’s have a look at where the principal accent lies on each word, the reason becomes clear:


‘Words’ then are accented units. The point is that for the inscriber of an inscription like ISic001231, a ‘word’ was an accented unit. Small words, in many of the world’s languages, tend not to carry their own accent, and instead rely on neighbouring words for their accentuation. (Incidentally, this is also why these kinds of words often get reduced in fast speech, e.g., ‘and’ in ‘fish ’n’ chips’.) This is why the small words ΚΑΙ “and” and ΣΥΝ “with” in ISic001231 don’t get their own word divider: they don’t carry their own accent.

What is the reason for ‘words’ in these inscriptions corresponding to accented units, rather than ‘dictionary’ words? Unfortunately, we can’t go back and ask. But the fact that ‘words’ are separated on this basis sits well with other things we know about how people interacted with written documents in the ancient Mediterranean world. The written word was much closer to the spoken word than it is for us: there was simply an expectation that the written word would be read aloud; even those who could read tended to read out loud to themselves, rather than silently, as we tend to do. Those who could not read would only have been able to access written documents in this way.

If interaction with the written word was primarily in the context of the spoken word, it makes complete sense to separate words on the basis of how you would say them out—rather like musical phrases—rather than on the basis of how you would look them up in a dictionary.

It turns out, then, that both ancient Northwest Semitic writing systems and Greek had in common the separation of words-as-they-were-pronounced rather than dictionary words. This keen interest in the oral component of what is written renders somewhat more mysterious the lack of vowel writing in Northwest Semitic writing systems: it isn’t the case that the writers simply weren’t interested in the oral component of writing. They were, as we see with how they placed their word dividers. They just happened not to show it by writing vowels.

You can download Rob’s monograph, The Semantics of word division in Northwest Semitic writing systems: Ugaritic, Phoenician, Hebrew, Moabite and Greek for free as an open access PDF here.

I.Sicily documents

ISic001231: Prag, J. R. W., Cummings, J., Chartrand, J., Vitale, V., Metcalfe, M., Antoniou, A. and Stoyanova, S. ‘I.Sicily 001231.’ Revised 2021-01-19.; doi: 10.5281/zenodo.4352237


Crellin, R. S. D. (forthcoming) “Word-level punctuation in Latin and Greek inscriptions from Sicily of the Imperial period” in Steele, P. M. and Boyes, P. (eds) Writing around the Mediterranean: Practices and adaptations. Oxford: Oxbow, 195–219.

Sampson, G. (2015) Writing Systems, 2nd ed. London: Equinox Publishing.

Waal, W. (2020) “Mother or sister? Rethinking the origins of the Greek alphabet and its relation to the other ‘western’ alphabets” in Boyes, P. J. and Steele, P. M. (eds) Understanding relations between scripts II: Early alphabets. Oxford: Oxbow, 111–124.

~ Robert Crellin (Former Research Associate on the CREWS Project)


One thought on “Dots between words in Northwest Semitic inscriptions

  1. Very interesting article and makes sense as to the written word meant to be spoken considering oral histories were the first language before the written text.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s