The evolutionnews.org blogpost contends that the analogy posited by the authors of the PNAS article breaks down because they assume that "functional proteins arose by natural selection," but while human languages may "evolve" in some sense, it is not by natural selection" because decision-making minds are involved, choosing how best to express thoughts." I would also argue that the analogy breaks down due to a debatable assertion made by the authors of the article and their misunderstanding of a basic fact of human language.
The debatable assertion appears when the authors of the PNAS article write: "Like all cellular life forms, all natural languages are believed to have descended from a single ancestor and have evolved through mechanisms comparable to biological evolution." The problem with this statement is that it is pure speculation. There is no actual evidence that human languages "have evolved through mechanisms comparable to biological evolution." If human languages evolved in a manner similar to what is posited for biological organisms, then we would expect to find that ancient languages were noticeably less complex than modern languages. However, that is not the case. For example, according to Irene Thompson at the aboutworldlanguages.com website, the Indian language Sanskrit is "one of the oldest, if not the oldest, [of] all attested human languages...The oldest form of Sanskrit is Vedic Sanskrit that dates back to the 2nd millennium BCE." Sanskrit is "a highly inflected language which uses prefixes, suffixes, infixes, and reduplication to form words and to represent grammatical categories." Sanskrit nouns have three genders (masculine, feminine, and neuter), three numbers (singular, dual, and plural), and eight cases (different forms of a noun which reflect its grammatical function, like subject or object). Sanskrit is in fact quite complex. As a further example, take a language that I personally have studied (to some extent)--Classical Chinese, which dates back more than two millennia. Classical Chinese, I can assure you, is no less complex than modern Chinese, which I have also studied. The idea that human language evolved from simplicity to complexity in the same way that biological organisms are thought to have evolved lacks any concrete proof.
The authors of the PNAS article also get a basic fact about human language wrong. They write that "in a written language, individual letters cannot carry semantic information; the smallest unit of information, therefore, is a word." No! The smallest unit of information in a language (written or spoken) is not a word, but a morpheme. A morpheme is the smallest unit of a language that carries meaning. All words are morphemes, but not all morphemes are words. A classic example would be the "s" that usually appears at the end of plural words in English. This "s" is not a word, but it does convey a meaning--"plural." Thus, a word like dogs contains two morphemes--dog and s. Furthermore, in written English, the single letter "a" can carry semantic information--it can mean "one."
As I mentioned at the beginning of this post, my understanding of biochemistry is quite limited, so I have no ability to assess the scientific merits of the argument made in the PNAS article. However, as someone with an academic background in linguistics, I can say that the authors of the article clearly stumbled in their effort to set up an analogy between human languages and proteins--which, I suppose, simply shows the danger of appropriating ideas from a field one has limited expertise in.
Image of a biochemistry laboratory from commons.wikimedia.org