*bumped*
Well, there is a problem with the term "celtiberian" that is still discussed mong the specialists. As you know, it's an exonymus coined by the romans. Strictly speaking, the "celtiberians" where a sort of confederation of tribes living in the Ebro valley and te iberian chain. But sometimes it is used as extension to all the "celtic" peoples in Iberia.
The first doubt about the term is what it is referrig exactly. That means celtic peoples mixed with iberians? That means celtic people living in Iberia? or tha means iberian peoples "celtified"?
On the other hand there were well known contacts between celtic (celtiberians, but also vettonii, vacceii etc peoples) and the iberian world. Even there are, apparently, linguistical shares among each others.
Well, I agree that the term "Celtiberian" is somewhat confusing, and I myself have used it somewhat confusingly.

First off, yes, "Celtiberi" in the strictest sense of the word only refers to the tribe of the Ebro valley. Still, other definitions are possible:
- the Roman exonym.
- quite similar to above, any Celtic-speaking peoples of the Iberian penninsula.
- Celtic-speaking peoples of the Iberian penninsula which have an Iberian background.
If find the latter two the most useful ones. The latter one in particular is worthy of note, because if we look at the situation in 200 BC, there is a clear-cut distinction between an apparently 'purely' Celtic area (in the northwest along the coast of the Atlantic) and an apparently 'purely' Iberian area (along the coast of the Mediterranean). Between that, you had an area of mixed affinities.
Regarding the general picture in the atlantic region. There are also evidences of a common cultural area in the region sice, at least, the bronze age, but probably before, as the bell-shaped culture(?) and the megalithics suggest.
The main problem I have with the general picture (and with any of the 'early' Celtization scenarios) is the existence of non-IE languages in the Iberian penninsula. It's easy to say that Celtic branched very early (ie, early Neolithic or even Mesolithic as Oppenheimer claims!) from the other IE languages and has been native to the Atlantic region since the Mesolithic - if you ignore the existence of other,
non-IE languages, in the region: around 200 BC, we have Aquitanian (Old Basque), Iberian and Turdetanian (which probably is the successor of Tartessian). The mere existence of these languages (and their diversity) casts doubt on any Meso-/Neolithic arrival of the Celtic languages in the Atlantic region. Conversely, there's a much stronger case for an ancient continuity of Basque/Aquitanian, both from the genetic and linguistic perspective:
- Haplogroup R1b has the highest concentration in the Atlantic region, and the Basques are almost entirely R1b.
- Basque is obviously an isolate language, wheras Celtic is part of the Indo-European language family. I admit that the situation with the Iberians and Tartessians is less clear-cut (mainly because both languages are extinct), because both definitely were heavily influenced by Greeks and Phoenicians by 200BC. However, the Tartessians were very definitely part of the Bronze Age Atlantic trade network - in fact their position can be explained by being part of both the Atlantic and the Mediterranean trade networks, thereby making Tartessos the 'hub' that connects both networks.
If we think this through (combined with the fact that Indo-European must have originated farther eastwards), we get that the Atlantic cultural area of the Neolithic (and probably well into the Bronze Age) was dominated by non-IE peoples.
In regard for the Beaker/Bell culture, they are late-neolithic through
earliest bronze age, and approximately contemporary to the Corded Ware culture. If we assume that Corded Ware was produced by a branch of early Indo-Europeans (which fits very well into the Kurgan hypothesis), then the Beaker/Bell culture was probably produced by non-IE peoples. The problem is, we know basically nothing about the ethnic identity of them - and we have no idea if they were something indigenous, or something else that established itself.
So, there is also problems with the celtic ethnogenesis in the area.
The Celtic ethnogenesis is a serious problem, and material cultures are of no reliable help here. Genetics isn't a help, either, here, because the genetic marker - Haplogroup R1a - which is commonly associated with the Indo-Europeans is largely absent from the Atlantic region, or arrived there only much later. The only thing we can assume from that without any reasonable doubt is that no large-scale population replacement took place in the region.
It seems very possible that the same lanes used by those atlantic peoples on close contact since old times would be used by the celtic cultural innovations to spread, whatever if it's a volkswanderung or only cultural transmision (I preffer the later). In that sense, there are rupture but also continuities in the "celtization" of the atlantic region. Not only in the reutilization of bronze age structures but also in certin dynamics. In the iberian case the castrum of Cogotas is a book exemple of that. There is a continuum since the Bronze Age until the celtic or "celtized" culture just before the roman conquest. So, maybe we should addopt a more open view about hat is a "celtic".
The only material culture that can be linked with Celts beyond any doubt are the Hallstatt and La-Tene cultures, however these represent only a part of the Celts, as the Irish and the Celtiberians (note: Celtiberians in the sense of any Celts on the Iberian penninsula

) were located clearly outside of Hallstatt/La-Tene influence for most of the time. The rapid expansion of La-Tene into the Atlantic region makes only sense if people in these areas already spoke a form of Celtic - and we indeed are talking about a cultural transmission then - just like you said.
Now, I had an idea there that perhaps linguistics can help here. There exists the idea of the split between Q-Celtic and P-Celtic, the former including Irish and Celtiberian, the latter including Gaulish, Bretonic and Welsh. It should be noted that the Q-Celtic languages are restricted then to the Atlantic region (Iberia and Ireland), wheras the P-Celtic languages are also found in the east. The eastward thrust of the Celts (which led to the invasion of Greece and the subsequent settlement of Celts in central Anatolia) seems to have been done exclusivel by P-Celts, as Galatian was probably very similar to Gaulish. Thande also mentioned that there's genetic evidence for the Irish link to Iberia.
Linguistics, genetics and even Irish legend (though the latter shouldn't be taken that seriously, it still gives us a hunch

) hold that the Irish have a relationship with the Celts of the Iberian peninsula.
If we look at the archaeological evidence, we must assume that Q-Celtic is probably the older branch of the two, and that P-Celtic was a linguistic innovation that occured farther in the east in the Hallstatt/La-Tene core area. At least, this hypothesis reconciles linguistic and archaeological evidence much better.
This in turn means that Q-Celtic must have arrived significantly earlier (not sure when, but probably in the Bronze Age), and perhaps we can interprete it that way that there were La-Tene "missionaries" who spread ironworking in the Atlantic region after circa 500BC. I admit that this doesn't solve the problem wholly, it only pushes things earlier.
Also, the problem with genetist as Oppenheimer, or Cavalli-Sforza for that matter, is that they have a tendence to hard-line difusionist interpretations, even ignoring the archaeologic evidence that points in other directions, often asimmilating genetics to culture. As other are saying, there is not neccesary causality between genetics and language. But also the achaeological cultures should be taken with caution. A common material culture or a common technic don't implies a common symbolic world. Of course that's an evidence of contacts ad cultural exchanging but beyod thatwe enter in the terrain of uncertainty...
Cheers.
Well, I agree that continuity of a material culture is by no means a proof of a common language.