Thanks to Tony Marsh for asking "Any thought from the etymologists among you on the rhyming of words:  CELERY and SALARY?" for inspiration for this.

The etymology given is the quintessential one so loved by antiquarians of Greek → Latin → French thence to poor English who couldn't invent their own words. So the name "celery" in the English language is thought to have been derived from the French word "celeri." In turn, celeri seems to have come from the Greek in which it first made an in Homer's Odyssey as "selinon". So, salary - salt - celery would appear to be linked because salary derives from Latin for Salt  (Sal) and at least some varieties of wild celery are salt tolerant plants.

However, I never beliieve such "etymologies" and I find it helps to look for one Anglo Saxon, because if I finds possible etymology in Anglo Saxon when it is supposed to be Latin it tells me:

  1. It's easy to make up etymologies (and/or)
  2. The antiquarians had a very dim view of Anglo Saxon and only ever looked there if they couldn't find some darned nonsense in Latin or Greek

So I look up the Anglo-Saxon etymology.

Here "sylu" means a miry place. Syla means a ploughman so I presume this "linguistic space" is the origin of soil and sully.

So, sel/sil/syl type words occupy the "linguistic real estate" of soil & silt.

Looking around we have seahl (willow) and sealt (salt) and modern silt (supposedly from a Scandinavian word related to sealt).

Unfortunately, the first attested mention of celery is too late (1660) to give any clear indication of its paternity. This can often mean that an Anglo-Saxon word used by ordinary folk has finally found its way into the hallowed cloisters of academia (i.e. its origin in the church) and a word used for centuries has finally been recorded.

But it can also mean the opposite - that a word hardly known outside acadamia/ the church has come into use in common English (such as democracy). But it is all far more complicated than that, because Latin was itself a mish mash of languages and many "Latin" words are undoubtedly of "barbarian" origin in the same way English is full of words from the empire from "pyjama" to "trek".

So many words will have followed the flow of "barbarian → Latin → French (e.g. car - with its antiquarian etymology: "Latin for a two wheeled Gallic cart" - I wonder what the Gauls called it?). And where those barbarians were British, the similarity of a Latin word may be due to assimilation in Latin rather than the reverse.

So, the only way to really know is to find out whether the word has "relatives" in one language or another. Obvious ones are "salt" and "salty". But consider the following. Both Silver (sealfor) and seal (seolh) are present in Anglo Saxon so are these rhyming words are related? And is it any accident that "Seal" (Seolh) starts with "Sea" (sae variant seon) - and what then "Silvery sea" and "Silkie seal". (Silk, AS: seoloc)

Common Ancestor or strange attractor?

The simple fact is that words that sound alike often have similar meanings. There are two causal explanations (the other is just chance). One is that there is divergence from a common "ancestor". And example of this given above is "sully" and "soil".

The other thing that tends to happen is that words with similar meanings grow alike. Take e.g. Cockatoo. The original in 1610 was from Malay Kakatua (possibly derived from its call). But over time the word has been "pulled" toward the English word "cock" to become cockatoo. Or "Fount" as in "fount of all knowledge". A shortening of "fountain", but thought to be influenced by "font". Or "Hokum" which seems to have obtained its form in mimic of "bunkum" and the phrase hocus-pocus. Then there is "nor", a contraction of middle english "nauther" which was influenced or "pulled" toward its final form by "or".

Roman Bath

Roman bath is recorded as having the name "Sulis". In deference to the antiquarian idea that welsh-speaking Britons were all massacred by the Anglo Saxon invaders in a mass genocide that left absolutely no archaeological evidence and is in my view total hokum, antiquarians searched in Welsh for a suitable "parent" and found none.

Now Bath is so called, because of the hot springs that in there natural form are often nothing more than steaming muddy puddles.

And so when one looks in Anglo-Saxon and finds "sylu" means a miry place, there is no doubt that "Sulis" would be readily understood by any Anglo Saxon speaker.


There is absolutely no evidence of a massacre of Welsh speaking Britons. There is almost no evidence of any place names of Welsh origin outside Wales except the name "Avon" which is readily derived from the Anglo Saxon word "Haefen", "*aeven" (haven or port) which applies to most rivers when they reach the sea). So there is no basis to fabricate stories of massacres of Welsh Britons nor believe that the majority of England did not have substantially the same Germanic based language at the time of the Romans as did did after the Anglo Saxon conquest.

So the most likely etymology of Bath is simply "mud".

Back to Celery

I started with celery and showed how it is related in sound and meaning to salt and silt and even sully and Sulis.

So is Celery Anglo Saxon or Latin? For that, language is not enough. Instead we need to look at both textual sources and biology.

What we are told is that celery is thought to have originally grown wild in the Mediterranean basin where texts from before 850BC describe its use for medicinal purposes. By the time, the Greek writer Homer referred to it the name was "selinon", from which either came or was dervied, the the Latin name: "selinun". When it arrives in French we find "celeri". So, the biological ancestry shows us this plant moving from the Mediterranean to Britain bringing its name to these shores.

So, just as it is ridiculous to suggest that the Latin name for a "Car" was not derived from the Gallic name for the same vehicle, so it would be churlish to believe the name did not come from the Mediterranean.

But then, what of "salt" or latin "sal". Could "Sal" have been an "intrusive" element - much like an invasion of an exotic species like Mink not indigenous, but soon making its home in "Old British". Could this intrusive element have become so assimilated that by the time we have any written records, it had evolved a number of forms related to the sea like: seal, silt or even silver?

We really do not know, but we can say it is unlikely unless there was time to see this change. Was there time for all these changes to have occurred in a mere 1000 years? And if not, then the fact that one language has many related forms is very strong evidence that the word originally came from that language rather than another where similar forms are poorly attested.

To use an analogy. If you found a lost child and two families claimed they were there's (or more likely denied having anything to do with the brat). Could you tell which family they came from by whether the family as a whole looked similar. Yes there will be uncles and aunts with no genetic similarity, but as a group, one would expect to find many closely related forms.

Likewise in the indigenous language for a word, we expect to find forms that by the process of "strange attraction" or divergence will be similar in both sound and meaning. And conversely, an intrusive word will have few close kin.

The origin of Celery

The available etymologies only take us back to Homer's Greek "Selinon". But that is more because that is one of the earliest texts we have, so we have no idea whether the word Selinon is a Greek word or one imported.

However if we look around at other languages we find three main other forms:

  1. Celeri (French), zelena (Slovene)
  2. Apio (Spanish), Accia (Sicilian), apium (Latin)
  3. Karafs (Arabic), Kereviz (Turkish), Karfusa (Maltese)

The first clearly derives from the ancient Greek σέλινον (sélinon). The Latin/Spanish form seems to be unrelated to the other two, but the last "Karafs" is interesting.

In English we are familiar with the hard and soft 'C' "Cat" and "Citizen". But other languages have a clearer distinction and the hard 'k' of Karafs is not necessarily the same as "Celeri". However, whilst ancient Greek "Selinion" is very different, the Mycenaean Greek has the form "se-ri-no" and "Ser-ino" and Arabic "Kar-afs" are too similar to easily dismiss as just coincidence.

They are even closer related if we look at Japanese where we find "serori". It is likely that the Japanese word stems from a late European word so it appears that the well known difficulty of Japanese speakers in pronouncing an 'l' has caused the change seleri  → serori.

And that is as far as I can go. The linguistic difference between "Celeri" and "Karafs" is too great to attribute to mere changes when words are assimilated. However the similarities mean they are linguistically close enough to make a single ancestor likely. So these appear to be two different "groups" of words within which we must assume the words come from one of two original words, but because the similarity of the two groups, it is reasonable to believe they came originally from the same word, even though we have no record of what word that is or even what language it came from.

However, we can also say that it is where we find the most similar form of the two families, that we would most likely find a linguistic missing-link that ties them together. So we can go on to say, that origin appears to be somewhere around the area of the Mycenaean area of influence where it overlaps with some language that influenced Persian, Arabic and Turkish.

Khoresh Karafs (Wikipedia)


I just happen to come across this page with a lot of Persian recipes including one for Khoresh Karafs (Celery Stew I suppose). Persian Khoresh (خورش‎) is a generic term for stew dishes in Persian cuisine.  There's an interesting similarity which hints (but no more) at a possible "linguistic family" of food related words in Persian or a related language.