What's wrong with using Pictish symbols as a justification for ID.

I know, I know. It's been forever since I've posted here. But I want to respond quickly and publicly to a post on Uncommon Descent, and they've banned me multiple times, so doing it there is a nonstarter. Briefly, UD regular PaV jumps for joy over a recent paper in Proc. Roy. Soc. A entitled "Pictish symbols revealed as a written language through application of Shannon entropy." PaV writes:

I’m not going to write too much. Just read this article and thump your head. If this isn’t an all-out admission of the validity of Dembski’s approach, then what is? I wonder if the Royal Society knew these authors were creationists? The article itself is open. Here it is.

BTW, the authors determine the Pictet symbols to be a language. As to the title of this thread, I consider languages to be designed. If you have a differing opinion, I would love to hear what it is!

Well now. Is this paper "an all-out admission of the validity of Dembski's approach"? PaV clarifies why he thinks so in a comment on the thread:

I don’t think that what these authors are writing about is anything new; in fact, they’re simply using methods that Shannon himself developed. The importance, however, seems to me to be that this is in a “peer-reviewed” journal, and that as scientists the authors are distinguishing between random images and images that have an underlying linguistic structure. This seems to me to be what ID argues. Thusly, if what these authors have done is acceptable science, then, so too, is the ID project.

Have the authors “proved” that these symbols are a language? No. Can ID “prove” that there is an Intelligent Agent responsible for life? No. But, in both cases, this is the best working hypothesis–or, in the language that Stephen Meyers uses, these hypotheses have the most “explanatory power”.

I think PaV is confused by the word "random." The paper's authors use the term in the abstract:

Using the technique on the Pictish symbols established that it is unlikely
that they are random or sematographic (heraldic) characters, but that they exhibit the characteristics of written languages.

But the authors do not mean "random" in the way that Dembski does, and the differences reveal why this is not a justification for ID. Remember, Dembski wants to distinguish between naturally occurring and designed objects or object features. He claims that we can do this purely from the object considered, without knowing anything about its history or the putative designer. In this way we are not, in Dembski's scheme, limited to human or even material designers of material objects.

With Pictish, however, we know:

1. The symbols are all artifactual (designed).
   
2. They are produced by humans.
   
3. They may or may not be expressions of written language.
   

The first sentence of the paper refers to "the durable artifacts left by primitive societies." Later, the authors describe the Pict stones as follows:

Picts did, however, leave a range of finely carved stones inscribed with glyphs of unknown meaning, known as ‘Pictish Symbol Stones’.

So the stones are designed, and they are designed by humans -- Picts, in particular -- who, like other Iron age humans, presumably used language. The point is not to distinguish designed from non-designed objects but linguistic from non-linguistic design.

Further, they are able to make this determination only because they know something about human language. They write:

This paper describes a technique that incorporates linguistic functions in order to quantify the level of communication in these small, ‘incomplete’ symbol datasets and thus differentiate between the different possible character types of writing (the term incomplete is used here to describe text samples that have insufficient data to properly characterize the character lexicons).

Sorry to repeat myself, but apparently this is lost on PaV: the study works by applying things we know about human language to determine whether these designed objects are human language. It takes information from a known domain (human language generally) and uses it to interpret and define a set of designed objects (Pict symbols: are they language or not?).

In short, the Pict study examines a difference of degree among objects known to be designed, and known to be designed by humans of a particular period. Dembski claims to be able to identify absolutely a difference in kind in situations where we have precisely zero information about any possible designer.