Consider the English sentence in (1): (1) I found a new app for my phone that I read about on a life hack blog. This sentence would be readily understood by many users of English in the early 2020s, but just 30 years earlier, it would have been very difficult for anyone to decode it, because some of the expressions used here did not exist or did not have the same meaning in the 1990s. In the following discussion and throughout the rest of this chapter, all etymological information for English is from the Oxford English Dictionary Online 2023, unless otherwise noted. For example, the word app as a short form of application had just started being used, but it was only used for computer programs. Smartphones were still very new, and app was not yet used to refer to programs for phones, since the very concept of installing programs on phones was not widespread. Online journals existed, but they wouldn’t be called blogs until the late 1990s. And the term hack had long been used to refer to infiltrating computer systems, but it would not be combined with life to refer to cleverly modifying one’s routine activities until the 2000s. If we go back even further in time, this sentence would essentially be uninterpretable. The sentence in (1) demonstrates one important pressure that can cause a language to change. The world around us is constantly changing, with new technology, cultural shifts, and growing scientific understanding, and we want to be able to talk about those changes. We could just use cumbersome descriptions, but the more common a concept is, the more likely we are to dedicate or adapt specialized terminology for it for more efficient communication. No one wants to keep saying clever modification to routine activities when they could get by with the much shorter expression life hack.
Changes in the world around us are not the only reasons that a language might change. As discussed in Section 2.5 and Chapter 10, languages naturally vary, and different linguistic variants can carry different social meanings. However, while variation may be stable over time (see Section 10.3), any given instance of sociolinguistic variation could eventually collapse, leaving older forms obsolete and forgotten, while the newer forms are now considered unremarkable rather than socially meaningful. For example, to describe the construction of a new house, speakers of English in the 1600s would have said either the house is building or the house is in building (as depicted in Figure 14.1, left). But by the 1700s, a new variant had emerged, with some speakers saying the house is being built instead (Figure 14.1, centre). This new variant was looked down upon by the educated elite at the time, who viewed it as “clumsy” and “at war with the genius of the English tongue” (Marsh 1862: 461–465). However, the newer form has now replaced the older forms, overcoming the prescriptive pressure against it, so that it no longer has any negative perception (Figure 14.1, right). Instead, the older forms are now the forms that would be out of place. Today, the house is building could only mean that the house is somehow building some other structure, while the house is in building just seems ungrammatical. Much of the discussion is this textbook has focused on synchronic linguistics, which is the study of language patterns at a single point in time, such as how the house is being built is grammatical right now. In this chapter, we turn to diachronic linguistics, which is the study of changes in language patterns across two or more points in time, such as how the house is in building used to be grammatical in English 500 years ago but is no longer grammatical today.
Sometimes Early and Late are also added before the name of a historical period of a language to make finer distinctions within that period when needed. For example, because Modern English spans almost 600 years, it is often useful to distinguish between Early Modern English (such as Shakespearean English) and Late Modern English (the English used in this textbook), given how different they are.
Note that this terminology is typically used only for attested languages, that is, languages that we have direct evidence of from writing or audiovisual recordings. See Sections 14.8, 14.9 and 14.10 for discussion of how we can deal with unattested languages for which there is no direct evidence.
There is no universal timeline for the use of this terminology. It is based on differences in the attested forms of the language, not by any particular year. For example, the earliest attestation of Old Chinese are engravings on ox bones and turtle shells that date back to 1200 BC (Boltz 1986), well before Old English.
In addition, there is normally no distinct single point in time at which a language strictly transitions from one period to the next. For example, Old English is sometimes considered to have ended in 1066, with the Norman conquest. However, the influence of Norman French on Old English was not instantaneous, and distinctly Old English linguistic patterns persisted until the 12th century. Thus, the transition from one period of a language’s history to another is like a smooth spectrum with no sharp divisions between periods, as shown in Figure 14.2 for English.
Furthermore, all languages naturally vary regardless of historical period, so there was no single consistent version of Middle English or Old Chinese or any other language. There was always some sort of variation based on class, age, gender, region, and/or other social factors. So when we talk about Middle English or Old Chinese as languages, we have to remember that they varied: the Middle English of royalty was different from the Middle English of farmers, the Old Chinese of western China was different from the Old Chinese of eastern China, and so on. Some of this variation is even attested in written records.
We may represent a language change with a greater than sign > pointing from the older form on the left to the newer form on the right. For example, the change in structure from early modern English the house is building to modern the house is being built could be represented as (2).
(2) the house is building > the house is being built
This notation can be used for any kind of change in any aspect of a language. For example, the pronunciation of the word prove used to be something like [prʌv] in Early Modern English, as needed for the rhyme with love in the final two lines of Shakespeare’s final sonnet, Sonnet 154 (Crystal 2011):
Came there for cure, and this by that I prove,
Love’s fire heats water, water cools not love.
The pronunciation of the vowel in prove eventually changed to [u], while the corresponding vowel in love did not, ruining Shakespeare’s perfect rhyme. We can notate this phonological change as [prʌv] > [pruv], with the older pronunciation on the left the greater than sign and the newer pronunciation on the right. Sometimes, a less than sign is used to indicate the reverse direction of a language change. Thus, we could write [pruv] < [prʌv] to indicate that the modern pronunciation [pruv] of prove originates from the older pronunciation [prʌv]. This means exactly the same thing as [prʌv] > [pruv].
Changes in meaning can similarly be represented with this notation. The word nice was borrowed into Middle English from Old French in the late 1200s, and its original meaning at the time in both languages was ‘silly’. Over time, its meaning changed in various ways, to ‘delicate’ (in the 1400s), then ‘careful’ (1500s), and eventually ‘pleasant’ (1700s). This overall sequence of change in meaning can be represented as (3).
(3) nice ‘silly’ > ‘delicate’ > ‘careful’ > ‘pleasant’
Additional information may sometimes be added to this notation, such as the name of the language or language variety at each stage. For example, to show a deeper history of the word nice, we could go back to the original Old French source and show its development from Latin, as in (4).
(4) English nice ‘pleasant’ < Old French nice ‘silly’ < Latin nescius ‘ignorant’
(5) Latin nescius ‘ignorant’ > Old French nice ‘silly’ > English nice ‘pleasant’
Note the similarity in usage between the greater than sign > and the arrow →→ for indicating a change in some linguistic structure. The difference is that the greater than sign > represents a diachronic change from one time period to another, while the arrow →→ represents a synchronic change at a single point in time, such as how a phonological rule converts an underlying representation into a surface representation.
The historical development of a word or expression is called its etymology. The etymology of nice is an example of how the modern meaning of a word can be very different from its etymological source. In this case, nice started with a negative meaning in English but eventually developed a positive meaning.
This is a reminder that a word’s meaning is defined by how the word is used now, regardless of how it may have been used 500 years ago, 50 years ago, or even 5 years ago. See also related discussion in Section 3.1 about the term phonetics and other scientific terminology. It is an etymological fallacy to make a logical argument that relies on using an etymological meaning of a word in a context where its current meaning is what is relevant.
For example, if someone is told that their behaviour is transphobic, they may try to derail the conversation and downplay the offensiveness of their actions by focusing on the meaning of transphobic instead. They might argue that they could not be transphobic because they are not literally afraid of trans people.
However, this is an etymological fallacy. The suffix –phobic is used in modern English to refer to being averse in a more general sense, not necessarily due to fear specifically. We see this modern meaning in various words referring to social intolerance or discrimination, such as transphobic, Islamophobic, and xenophobic, as well as other words referring to physical incompatibility or intolerance, such as hydrophobic (which refers to being water-repellant) and photophobic (which refers to being physically sensitive to light).
There are two broad categories of sources of language change, external and internal. External change is change caused by something from outside the language itself, such as cultural changes, new technology, etc. An important external source of language change is language contact, when two or more cultures who use different languages interact with each other, which results in influence between the two languages. There are many different ways that cultures may interact, some more egalitarian than others, and there are consequently many different ways that language contact can cause language change. Some of these are discussed in Section 14.7.
But in many cases, a language may change for no external reason, which results in an internal change that happens spontaneously within the language due to its structure or modality. For example, when two phones are adjacent, it is often easier to pronounce them by assimilating some phonetic property from one to the other, so that they are more similar (see discussion of assimilation in Section 4.9). There is no external reason from the surrounding culture that causes assimilation. Assimilation is just something that is physically easier for human bodies to do, so this is a kind of change that could happen to any language at any time.
Many internal changes likely happen during the acquisition process. The earliest claim of this idea is probably Paul 1880, and there has been much work on this topic since, such as Ohala 1981, Crain et al. 2006, and Niyogi and Berwick 2009. During acquisition, children may articulate the language in a slightly different way than their parents. The change may be subtle enough that no one really picks up on it consciously. However, the change could still be robust enough that it persists into the next generation, who themselves may then push it just a bit further, and so on, for many generations. Although ordinary language users may not notice these subtle, incremental generational changes, linguists are often able to measure them in real time with modern instruments.
Many people believe that ancient humans must have used simpler language, and that modern languages are more complex. But many people also believe that younger generations are ruining language by simplifying it through a combination of laziness and ignorance. These contradictory views may even be held at the same time by the same person. It is probably no coincidence that many people believe that their own personal version of a language happens to be the best version.
In truth, language complexity and language change are not actually related. In part, this is because language complexity is not a single well-defined concept. Languages may be complex in one way but simple in another. For example, English has fairly simple verb morphology, with most verbs having only a small number of distinct forms (for example, sign, signs, signed, signing), but it has a complex vowel system with over a dozen vowels, including multiple diphthongs. In contrast, Georgian (a Karto-Zan language of the Kartvelian family, spoken in the Republic of Georgia) has only five vowels (Shosted and Chikovani 2006), but it has incredibly complex verb morphology, with as many as a dozen positions that can be filled by many different morphemes marking verb tense, agreement, etc. (Makharoblidze and Leonard 2022).
In addition, each aspect of a language can change in different ways. We find that some changes may reduce complexity, some may increase complexity, and some may have no obvious effect on complexity at all. Furthermore, while a change could continue along the same path, growing increasingly complex over time, it might instead reverse at any time, and even reverse again and again and again. Thus, there simply is no universal relationship between language change and complexity.
