So where does the OED leave us? Ah, yes. The Dictionary traces silver back to Proto-Germanic *silubra-. (The asterisk indicates that the word was never directly recorded, but that it was reconstructed by comparing the Germanic languages and their lexemes). The Dictionary also gives us a number of cognates, to which I will add even more via Boutkan and Siebinga (2013).
Old English: seolfur
Old Saxon: siluƀar
Old Norse: silf-r
Old High Germanic: sil(a)bar
Old Church Slavonic: sĭrebro
Old Prussian: siraplis
So that's a pretty fair list. From the get-go, we can see that whatever "proto" word we obtain will have a form akin to silub-r, silab-r, sileb-r or the ilk. The vowels are uncertain but we have a short list to choose from.
Now, where did the Germanic and Balto-Slavic tribes get this word? When most European tribes gave a name to the new metal, they picked a word from their native inventory. Not surprisingly, they usually called silver "the white (thing)," based on the Indo-European root *h2erǵ- "to shine," "white." That was true for the Celts (PCeltic *arganto-), the Greeks (Mycenaean a-ku-ro), the Armenians (PArmenian *arcant-), and the Italics (PItalic *argento-).
Because both the Balto-Slavs and the Germans have a very similar word for silver, it is suggestive that they borrowed the word from a tribe that traded silver with them. But whence? Three telling hints come to us from the Iberian peninsula.
1. Basque zilar < Proto-Basque *zirar. Löpelmann (1968) believed the word to be a loan from Germanic, but the famous Vasconist, R. L. Trask, notes that the direction of German-to-Basque could not be firmly established (2008). Further, what if it was a third-party loan into German and Basque? Nevertheless, the word gives us a firm relationship to the Iberian peninsula.
2. Celtiberrian silapur. Celtiberrian was a Celtic language spoken in the center of Iberia. This word cannot be traced back to a Proto-Celtic form, rather it stands alone in the Celtic family, again indicating a loan from an unknown source (Matasović 2013).
3. Iberian sarif, salir. Could this be the source-tongue we were searching for? Maybe. Alone this example stands as a word that was not necessarily borrowed. Note that iltiŕta-śalir-ban appears on coins. Could śalir-ban be our quarry? It is not a stretch to see the morpheme -ban misinterpreted by listeners as salirba-, a single word. Because ban also meant the numeral 'one,' could the word have been used in Iberian in the same way Basque uses their numeral bat 'one?' The Basque can utilize bat as a definite article which appears in the same position of a sentence as Iberian -ban has appeared. Thus the phrase may mean "an iltiŕta silver coin." On the other hand, the vowels are metathesized from an i-a pattern, a pattern we see in all the other languages, to a-i. So perhaps we finally found the donor language (Polome 1989), but in my opinion the jury is still out. It's good to see, however, the complete tale available in one page online somewhere.
- Boutkan, Dirk and Sjoerd Michiel Siebinga. "selover" in: Old Frisian Etymological Dictionary. Indo-European Etymological Dictionaries Online. Edited by Alexander Lubotsky. Brill, 2013. Brill Online. May 8, 2013.
- Löpelmann, Martin. Etymologisches Wörterbuch der Baskischen Sprache. Berlin: De Gruyter. 1968.
- Matasović, Ranko. "arganto-" in: Etymological Dictionary of Proto-Celtic. Indo-European Etymological Dictionaries Online. Edited by Alexander Lubotsky. Brill, 2013. Brill Online. May 8, 2013.
- Polome, Edgar C. “Preparing an Etymological Dictionary of Proto-Germanic.” Journal of the Dictionary Society of North America. No. 11. 1989.
- Trask, R. L. "zilar" in: The Etymological Dictionary of Basque. University of Sussex. Edited by Max Wheeler. 2008.