Charles Darwin’s theory of evolution by means of natural selection remains sturdily in place as the best explanation for living organisms’ manifold complexities. It’s so simple: during the process of reproduction, minor genetic variations that are beneficial to the survival of an organism are more likely to be transmitted to subsequent offspring. Due to the benefit of this variation, more and more of these modified offspring may survive, and the inherited trait might come to hold sway.

Natural selection is such a compelling explanation that, in the 150 years since it was first proposed, serious scholarship has yet to formulate even a single alternative theory. By virtually any extant scholarly standard, Darwin’s idea appears indestructible.

Alas, the overwhelming power of Darwinism has seduced the muddy-headed into some infamous misapplications of the theory (for example, in economics, social policy, and urban planning). But metaphorical Darwinism should not be dismissed out of hand even though a few unsavory sorts might misappropriate its tenets. There are cases for which the theory of evolution by means of natural selection is a wholly appropriate metaphor. The structure of language is one such case.

When we speak, we do not reproduce a perfect replica of the speech around us. Instead, there is limited variation—in the sounds we make, in the words we use—that is inherent to our patterns of speech. Consequently, though speech patterns are sufficiently stable to fulfill their communicative function, they are also sufficiently variable to be under constant—if slow-going—modification. Much as in genuine natural selection, systemic changes in language are often a consequence of the communicative success or failure of the individual word variants that we use. Successful speech propagates: today’s spontaneous, unplanned innovation may become tomorrow’s new norm, a norm that, utterly naturally and without a guiding hand, fulfills the communicative function of language.

Take a simple example. In most regions of the United States, the words “pin” and “pen” are pronounced sufficiently differently so as not be confused with each other. However, in the American South, the sequences ‘-en’ and ‘-in’ have evolved to a state such that both words—“pin” and “pen”—now sound the same (much like “pin”). There are complex though well-understood reasons for this potential confusion, having to do with “the four ‘A’s’”: acoustics (sound waves), articulation (our speech apparatus), aerodynamics (the flow of air), and audition (our hearing mechanism). From a communicative standpoint, this is potentially bad news, because if distinctions among words are eliminated en masse, then communication may become genuinely eroded.

But even the most linguistically unsophisticated among us will quickly realize that such an eroded linguistic state could never develop in English, or in any other language, for that matter: languages are always structured such that communication proceeds unencumbered. In the case at hand, just as “pin” and “pen” were merging toward each other in the south, a new compound word began to emerge: “pen” was replaced with “inkpen”. Lo and behold, the words that we feared—for one brief moment—might be rendered confusable, have emerged as distinct from one another, and so communication proceeds unhindered.

Although it is very tempting to believe otherwise, this avoidance of miscommunication did not happen by intention or design. There was no shape-changing plan acting on southern English which promoted the use of the word “inkpen” just at the point in time when ‘-in’ and ‘-en’ were becoming confusable. (“Inkpen” surely did not arise as an innovation in writing either, since the written forms “pin” and “pen” are contentedly non-identical.)

Instead, there are natural, passive pressures on the words we say—due to their everyday use and disuse—that resulted in the evolution of this compound word. Certain minor, chance variations in usage that just happen to render words less ambiguous—such as a sporadic southern speaker, here and there, using “inkpen” instead of “pen”—are more successful at communicating a speaker’s intended meaning. Due to this communicative success, listeners, having clearly heard the words and understood their associated meaning, are—passively and unconsciously—more likely to reproduce this usage in their own speech. From small things, big things come: beneficial variants propagate, and so an unplanned innovation may gradually become a new norm.

By contrast, mis-communicated words are more likely to passively fall by the wayside, perhaps, over time, being eliminated entirely: since they are sporadically misunderstood, they are—again, passively and unconsciously— less and less likely to be reproduced.

So, as “pin” and “pen” began to merge toward each other, the “inkpen” innovation served a dual role. First, it rendered the words clearly distinct from each other. Second, it may have actually sped the decay of the “pin”-“pen” distinction: since “ink” came to the rescue, there were fewer functional pressures that would promote the continued maintenance of any “-in” versus “-en” distinction.

Of course, the change from “pen” to “inkpen” only involves a single word, and so I have hardly demonstrated my claim that genuine systemic modifications can take hold in a language due to minor spontaneous variations in speech. But compelling evidence for full-blown change may be marshaled with ease.

For example, the Chinese of long ago had many single-syllable words that ended in ‘p’, ‘t’, ‘k’, ‘m’, ‘n’, or  ‘ng’ (in English, we use the digraph ‘ng’ to represent a single sound). Some contemporary dialects like Cantonese (spoken in southeast China) retain these six consonants. But others, such as Mandarin, have drastically reduced this set to only two members, “n”, and “ng”. In general, when consonants are not immediately followed by vowels (for example, at the end of a syllable, or at the end of a word [most words start with consonants, for well-understood auditory reasons]), they have a creeping tendency to be misheard, or not heard at all. If these consonants eventually wither away, then many words may be rendered indistinguishable from each other. This is what seems to have happened in Mandarin.

Does this mean that Mandarin has fewer syllable shapes than Cantonese does? Yes, it does. Cantonese has about 1800 syllable shapes, but Mandarin has only about 1300. Does this mean that Mandarin speakers are constantly misunderstanding each other? No, of course not. Mandarin, unlike Cantonese, has evolved a huge inventory of two-element compounds, which means that its words are now usually two syllables in length, and so have ample opportunity to maintain phonetic distinctness amongst themselves. The slow-going loss of the final consonants has been offset by the co-evolution of a compounding process.

In Mandarin, individual words at the risk of ambiguity—words that were identical except for their final consonants—may have been the first to evolve into two-syllable compounds, but these pioneering words may have eventually triggered a huge overhaul of Mandarin word structure: once a pattern is set in motion due to some functional pressure, we humans have a demonstrated tendency to generalize the pattern to new cases, such that it may come to hold sway. The result is that the functional origins of the pattern may be partially obscured by subsequent developments. Is this any different from genuine evolution? Sometimes, evolutionary biologists puzzle over the functional benefit of one characteristic or another. These scientists know, however, that their puzzlement is not due to a flaw in the theory of natural selection, but is instead due to their inevitably imperfect knowledge of historical developments.

Despite the fact that word and sound-patterns of languages display remarkable complexity, with demonstrable pull-and-tug interdependencies suggesting that their structural integrity is no more secure than a house of cards, languages are, in fact, always at the ready in a lock-and-load state: they are always, unrelentingly, prepared to fulfill their job of encoding and transmitting meaning from speakers to listeners.

Darwinism, indestructible though it may be, must be protected against those who would exploit its principles toward ethically unsavory or scientifically unsound ends. But, with caution and prudence, Darwinian thinking can and should be freed from its original confines. Genuine Darwinism cannot account for the structural properties of human languages. But the southern “inkpen” innovation, along with the countless other comparable cases found in linguistic evolution, show that the Darwinian metaphor may be embraced as a genuine explanation for the structure of our species’ remarkably complex system of communication.