Historical linguistics is the sub field of linguistics that concerns itself with language change. Thus far in the course, we've looked at the core areas of syntax, morphology, semantics, phonetics, and phonology. Historical linguistics touches on all of these areas, given that all of them figure into how languages change. In that sense, in order to study historical linguistics, it is important to first have a solid grounding in the general core components of grammatical theory.

One major difference, however, between historical linguistics and what we've done so far is that we have been analyzing language synchronically. This is a fancy way of saying that we've been looking at language at a particular point in time. By contrast, historical linguists look at language diachronically, which is to say that they look at language development through time.

• synchronic linguistics: the study of natural language grammar at a given point in time (e.g. the present)
• diachronic linguistics: the study of natural language grammar throughout time (i.e. the study of language change/evolution through time)

• What kinds of changes do languages exhibit, and why?
• What kinds of changes don't languages exhibit, and why?
• How are different languages related to each other, and how can we develop a dependable methodology for establishing these relationships?

A major turning point in the way that we view relationships among different languages came about two centuries ago, when an English guy named Sir William Jones noticed striking similarities between Sanskrit, ancient Greek, and Latin. The major breakthrough basically revolved around the following reasoning: the best way that we can account for the enormous degree of similarity is to assume that all three of these languages are descended from a common source language--a language that linguists have come to call Proto-Indo-European.

The notion that these languages had a common origin lent itself nicely to viewing language relationships in terms of a genealogical or "family" tree. The basic assumption behind the family tree view of language change is that languages change in systematic ways over time. This idea was put forth by August Schleicher in the late 19th Century and is known as the regularity hypothesis. In broad strokes, the reason that the regularity hypothesis is particularly important is that it allows us to compare different languages and to reconstruct their common ancestor. If language change were simply chaotic, we wouldn't know what we were comparing, and our reconstruction would be difficult. This whole operation of comparing languages and reconstructing a common source (if the evidence suggests that there is one) is known as the comparative method.

Remember the terms we used to label our trees in syntax? We use the same kinds of labels in historical family trees. So, if you look a the family tree diagram of Indo-Eurpoean on p. 381 of Language Files (8th ed.), you should know that the way to read such a tree is to view any languages that share a common source immediately above them as sisters. The language above them is called the mother. So, with respect to the mother language, they are daughters. So, Baltic and Slavic are sisters, and these two sisters are daughters of Balto-Slavic. For its part, Balto-Slavic is sister to Iranian, Germanic, Celtic, Italic and a number of other Proto-languages, all of whom are in turn daughters of Proto-Indo-European. In this way, family trees give a very clear picture of the direction of language change through time, as well as of how closely related two languages are (think first cousins, second cousins, and so on).

This said, it is useful to bear in mind that the tree diagram, while very useful, has a number of drawbacks that should be recognized. Your book lists a two big ones, which I'll highlight here:

• Trees give the impression of uniformity within a language, while the reality is that languages are comprised of a great deal of variation; in particular, they are collections of dialects which may vary in rather marked ways. Family tree diagrams make languages look more uniform than they are.
• Similarly, trees make languages look as though they split off very sharply from one another and never again entered into close contact, obscuring the fact that languages grow apart along various paths and in stages and that they may maintain a high degree of contact after their split. Languages grow apart in increments, not all at once. Family tree diagrams make the situation look different. It is perhaps more useful to think about different dialects within a language gradually becoming different languages over long stretches of time.

A concrete example of how this works is the following. If you look at the picture on p. 380, you'll see that Baltic and Slavic share an important feature with Germanic. These three are enclosed in a loop. What this loop actually refers to is that these languages share morphological plural case endings that have an [m] versus an aspirated [bh] is found in many other Indo-Eurpoean languages historically. At the same time, however, Baltic and Slavic (but not German) are grouped with Indic and Iranian by another loop. This loop actually refers to an extensive amount of palatalization (a phonological property) that is particular to these but not other languages in the Indo-European family.

What emerges is a much more finely grained (and messier) picture of relationships between languages. Note, however, that like the tree model, wave theory has its own drawbacks. Here's a couple:

• For one thing, lots of languages have similarities, but this isn't a guarantee that they are related. How can such similarities arise? Well, numerous ways. They might come about through borrowing when languages are in contact. You wouldn't want to say that English and Spanish are related because the words "adios amigo" and "fiesta" have been borrowed into contemporary English (though there are other good reasons to establish a relationship between English and Spanish!).
• For another thing, similarities might also arise by accident, much as convergent evolution may make two animals appear to be more closely related than they are. The term for this is language drift.
• Finally, the wave model doesn't give a clear picture of the direction of diachronic change, i.e. of which languages are sisters of a single mother, etc...

Okay, so languages change over time. How? Here, we'll look at a number of dimensions of change, beginning with the one that we know the most about and which consequently receives the most attention: sound change. (Just when you thought you were free from phonology!!!!!)

Top five reasons to study sound change:

• linguists have been doing it for a long time and we know a lot about it
• sound change often is crucial to our understanding aspects of morphological and syntactic change
• sound change is the best device for reconstructing the words of proto-languages
• sound change provides a good example of the concerns and methods of the historical linguist
• sound change is just plain fun!

So, we have a sound change in which Old English /k/ became [ch] before the high, long, front vowel /i/. There is one thing to bear in mind, though. Just because there is a phonological rule at work, doesn't mean that we have a historical sound change. Your book makes a point of noting this, and they are right to do so. They cite the example of present day English. If you think about it, you'll probably note that when you say the word "interesting", you sometimes delete the schwa that comes between the [t] and [r], resulting in something like "intresting". This is a rule of unstressed schwa deletion between [t] and [r]. But, you'll probably also notice that you don't ALWAYS delete schwa in this context, i.e. that there are times when you pronounce the schwa, especially if you pay attention to your speech in less casual circumstances.

Why is this important? For our purposes, it's an example of a phonological rule (schwa deletion) that does not constitute a wholesale sound change. Note, however, that if over time speakers begin to always delete schwa between [t] and [r] in all circumstances, then our phonological rule will have yielded a diachronic sound change: where there used to be schwa, there no longer is. Such a case is the rule of palatalization that your book discusses from Old English. EVERYWHERE that [k] preceded [i:] at an earlier stage in the Old English period, by the late Old English period, all of those [k]'s were being pronounced as the affricate [ch]. Without exception. That's a sound change.

Sound changes come in a number of flavors. Historical linguists make two broad distinctions. These are: conditioned vs. unconditioned sound changes. We use the term conditioned when there is something that can be said to have caused a particular sound change. In the case of palatalization in Old English, where [k] became [ch] before [i:], we say that this change is conditioned, because it is the presence of [i:] that is driving the rule. In contrast, unconditioned changes can be thought of as arising spontaneously. That is, we cannot attribute the sound change to something in the environment of the sound that has undergone the change.

Within these categories, there are a number of subdivisions--kind of like Cary, North Carolina! Briefly, I'll review these here.

Types of conditioned changes:

Assimilation: assimilation refers to a sound change that results in the affected sound becoming more like some sound or sounds around it. Think chameleon here. Assimilation involves changing to be more like your environment. An example in the history of English is a sound change in Old English in which [f] become voiced (i.e. it turned into [v]) between vowels. So, the word [wolfas] became [wolvas] 'wolves'. Since vowels are voiced sounds, this rule is assimilatory. That is, [f], be becoming voiced and turning into [v], became more like the sounds that surrounded it.

Dissimilation: If assimilation is about fitting in, then dissimilation is about standing out. Dissimilation is a situation in which a sound changes to become less like a sound or sounds in its immediate environment. In some varieties of modern English, for example, the final interdental fricative in the word 'fifth' is systematically pronounced as a [t], as in [fIft]. Here, we have a situation in which a fricative becomes a stop after a preceding fricative. That is, what was an f+th combination becomes f+t. This is dissimilation in that [t], a stop, is less like the fricative [f] than the fricative [th] was. The idea is that it is difficult to make two consecutive yet different fricatives. So, one of the two becomes a stop, thus simplifying things in articulatory terms. In this case, though, simplification is achieved through dissimilation--i.e. through making one sound less like a neighbor.

Metathesis: Metathesis means that two sounds change the order in which they appear in a word. In some varieties of English we have a metathesis rule that has switched the order of [s] and [k] when they appear at the end of a syllable, as in the word [aks] instead of [ask]. A simple way of understanding how metathesis works is this: xy--> yx.

Deletion/Insertion: Sometimes, sounds are deleted or inserted in particular contexts. For example, at the end of the Middle English period, unstressed schwas were deleted at the end of words. That is, in fact, the reason why so many words are spelled with a final "e" although there's no vowel actually there, as in the words "nose". There was, historically, an unstressed schwa at the end of the word. Anyway, such changes are "conditioned" in that it wasn't simply an across-the-board purging of schwas, but rather, there was a deletion rule that only affect schwas of a particular type (unstressed) in a particular context (word-finally).
 
 

Types of unconditioned changes:

Monopthongization: Monophthongization refers to a sound change in which diphthongs become monophthongs. OF COURSE, you need to know what diphthongs and monophthongs are to understand this! A diphthong is a complex vowel, i.e. a vowel that is a combination of two different vowels, as in the words "house" or "ouch" which phonetically are pronounced with an [a] vowel that is bundled together with an immediately following glide [w]. A monophthong is just a fancy way of talking about a regular vowel, i.e. a vowel that isn't a diphthong, as in the simple [a] in "father".

In so-called standard American English, the pronoun that we spell "I" is pronounced as a diphthong: [ay]. This is the way that I pronounce this word, for example. However, there are many dialects of English, including Southern dialects of American English, in which the diphthong [ay] has become monophthongized to [a:]. That is, the glide part of the diphthong is lost. Some if not many of you may in fact pronounce the word "I" this way. The reason that such a change is called unconditioned is that there is nothing in the environment that causes monophthongization. It just happens to some diphthongs wherever they are.

Diphthongization: Well, if you get monophthongization, this one should be easy. It's just the opposite. Here, monophthongs become diphthongs. In Middle English, long [u:] became a diphthong [aw]. The effects of this change are found in words like "house". In Middle English, this word was pronounced [hu:s]. Now, of course, it's [haws], with a diphthong.

Raising/Lowering: Raising and lowering refer to historical changes in VOWEL HEIGHT. Needless to say, I'm going to hammer away again at the importance of understanding the phonetic dimensions of vowel height to understanding how raising and lowering changes work in historical linguistics. A famous case of raising in the history of English involves the raising of long [o:] to [u:] in Middle English. The word that we now pronounce as [nun] 'noon' is historically descended from [no:n]. If you remember your phonetic features for vowels, you'll recall that [o] is a mid vowel and [u] is a high vowel. So, a change from [o] to [u] is called raising.

Backing/Fronting: Just as raising and lowering refer to historical changes in vowel height, back and fronting refer to changes in how far forward or how far back in the mouth a vowel is produced. An historical rule of English that involved fronting is a change in which the back vowel [a] fronted to [ae]. The effects of the change are found in modern English words like "calf", "ask", and "path", all of which have and [ae] vowel now but which used to be pronounced with an [a]. Why is this called "fronting"? Again, you need to understand your phonetic features for vowels. The vowel [ae] is a low front vowel, while the vowel [a] is a low back vowel. So, a change from [a] to [ae] involves moving forward in the mouth, i.e. fronting.

We've just reviewed the kinds of sound change that arise over time. One more important point is to consider the effect of particular sound changes on the phonological system of the language in question.

When we say "effect", what we are referring to is the relationship of the change to the system of phonemes that the language has. In this regard, historical linguists talk about two types of change: phonetic vs. phonemic. Let's look at each and compare them.

Phonetic Change

In an obvious sense, all sound change is phonetic in that it involves changes in the production of sounds. But that's not what historical linguists mean when they talk about phonetic change. What they refer to is something that is a bit more complex. Specifically, they are referring to a change that doesn't affect the phonemic contrasts of the language, that is, a change that leaves the system of phonemes alone and only affects the allophones of the phonemes that are already present in the language. As usual, a concrete example will make the point more clearly.

In Middle English, the voiceless stop phonemes /p/, /t/, and /k/ were never aspirated. So, simplifying a bit for our purposes here, we can say that they only had plain stops as their allophones. From what we know from phonology, we can diagram this situation as follows:
 
 

/p/ /t/ /k/ (phonemes)
 
 

[p] [t] [k] (allophones)
 
 

Then, English gained a phonological rule of aspiration that resulted in the addition of aspirated allophones of these three phonemes. You all should know this from our discussion of phonology. We can see that this rule of aspiration is still rigorously applied in English today. Okay, so, what happens to our picture as a result of the aspiration rule? Well, the phonemes stay the same, but each phoneme has gained a new allophone:

/p/ /t/ /k/ (phonemes)
 
 

[ph] [p] [th] [t] [kh] [k] (allophones)

(note: the [h] symbols indicating aspiration should be raised, i.e. superscritped)

What's important to understand here is that the sound change introduced a new set of sounds into English: aspirated voiceless stops. BUT, the change DID NOT affect the phonemes. That is, it did not result in a change in the set of stop phonemes in English. We still have the same three phonemes /p/, /t/, and /k/ after the sound change. This is an example of a phonetic sound change. Here, the term phonetic means that the sound change did not cause a change in the set of phonemes in the language. Instead, it affected the allophones of the phonemes that were already there.

By the way, here's a question. Is aspiration a conditioned or an unconditioned change? The answer is that it is a conditioned change. Why? Because aspiration didn't just happen to all voiceless stops. Rather, it happened to voiceless stops in a particular context. In particular, as we know, only voiceless stops that are the first sound in a stressed syllable are aspirated. That is, there is a particular context that conditions the change.

Phonemic change

Phonemic change occurs when a sound change does result in a shake up of the phonemic inventory of a language. Here's a case in point. Old English did not have a /v/ phoneme. It DID have an /f/ however, and then a phonological rule arose in Old English by which /f/ became [z] between vowels. This happened in about 700 AD--a long time ago. At this point, the effect of the rule was just like the aspiration case above. That is, the voicing of /f/ between vowels resulted in the addition of a [v] allophone:

/f/ (phoneme)

/\

[f] [v] (allophones)

Subsequently, however, English borrowed a whole lot of words from French. Some of these words had [v]'s in them, and these [v] sounds weren't just limited to appearing between vowels. Well, as you can imagine, things can get a bit confusing. Your book gives you an example of what happened by comparing the words "safe" [sef] and "save" [sev]. Essentially, what happened was that the French borrowings resulted in a reanalysis of the English phonemic system. Instead of [v] being simply an allophone of /f/, /v/ became a phoneme in its own right:

/f/ /v/ (phonemes)
 
 

[f] [v] (allophones)
 
 

So, first there was a phonetic change where /f/ gained a [v] allophone. Then there was a great deal of borrowing from French in which other words with /v/ were brought into English, giving [v] a wider distribution than it had previously had and creating minimal pairs between /f/ and /v/. This resulted in the reorganization of the phonemic system. That is, /v/ became a new phoneme of English, rather than an allophone of /f/. This, then, is a case of phonemic change.

If sound change is about changes in the phonological system of a language, then you can probably guess what morphological change is all about. Yep, changes in the morphological system of a language. Your book mentions some things that might sound a bit scary at first--terms scuh as "proportional analogy" and "pardigm leveling"--and in this section of the review sheet, I hope to take you through some of the basic issues in morphological change, while making these concepts seem as simple and straightforward as they really are in big picture terms.

Analogy

So, let's start with the following concept: analogy. Independent of grammar, what's an analogy? An analogy is a kind of relation based on some kind of correspondence. SO, if I say to you, "happy is to sad as tall is to X," you can probably guess that X is "short". Basically, you did this by analogy, because you looked at the relationship between "happy" and "sad" and realized it was one of antonymy (oppositeness--remember semantics!!!???), and by analogy, you supplied a word that was an antonym of "tall". This kind of thing shows up to torture victims on SAT exams every year. Anyway, I think (hope) that we all have a pretty clear sense that filling in a word such as "automobile" would be a strange answer for X: "happy is to sad as tall is to automobile." If so, then we also have an intuitive handle on analogy.

Analogy, as it turns out, is a major player in morphological change. Here's an example from Language Files. In Middle English, the past tense of the verb climb was /klom/. This is one of a number of verbs that formed its past tense by changing the vowel in the verb root rather than by suffixing [-ed]. Today, of course, we say [klaymd] "climbed". Clearly, we've regularized the past tense of this verb to make it more like the past tense forms of other verbs of English. In this sense, we can say that the change in the past tense of "climb" is attributable to analogy. So, rhyme is to rhym+ed as climb is to climb+X. Here, we plug in -ed for X and can see how the analogy works. We get climbed. Your book calls this kind of analogy proportional analogy.

When analogy (or analogical change) is used in this way, you can see that it results in the regularization of the morphological system. Think for a moment about the plural system in English. It is highly regular. To make a word plural, you add the plural suffix [-s] (actually, there are some phonological rules here, but we'll skip over these). Anyway, although the system is highly regular, there are some irregular plurals such as oxen and children or lice and mice, which reflect other ways that English also used to make plurals. Over time, by analogy, many words simply changed over to the [-s] suffix, leaving only a handful of words as relics of earlier stages of English morphology. Generally speaking, this kind of regularizing analogy results from there being a critical mass of words that are able to influence other words--kind of like morphological peer pressure!

Paradigm Leveling

Paradigm leveling is an interesting case of analogical change that makes a system more regular. Needless to say, we need to understand what a paradigm is to understand paradigm leveling. I'm going to give you a simple, working definition here. A paradigm is a group of forms that are related by inflectional morphology. A really simple example is the person inflections that we find for Spanish verbs. Here's an example of the present tense paradigm for the Spanish verb vivir 'to live'.

VIVIR

viv-o 'I live' viv-imos 'we live'

viv-es 'you live' viv-is 'you pl. live'

viv-e 's/he, it, lives' viv-en 'they live'

'you (formal) live' 'you (formal-plural) live'
 
 

Though I am admittedly simplifying somewhat, you can see that the verb root is /viv-/ and the various person/tense suffixes are -o, -es, -e, -imos, -is, -en.

So, what is paradigm leveling? Paradigm leveling is simply analogical change that makes the members of a paradigm look more like one another. It's then a case of regularization within a paradigm. Your book discusses a case of paradigm leveling in Latin that is a nice example.

Latin had a bunch of case endings such as accusative (object) and genetive (possessive) suffixes that went on the end of words. So, the nominative root honos 'honor' was honos-is in the genetive and honos-em in the accusative. What's important is that the root was always the same and then the suffix was added. The paradigm was perfectly regular.

OK, that's easy. Then, along came our friend phonology! Phonology plays a role in this discussion because Latin underwent a rule of sound change in which at a later stage of Latin, [s] became [r] when it showed up between vowels. So, what happens when we add the genetive suffix -is to the root honos? Yep, you guessed it. We created a context in which the final [s] of the root is between two vowels, and so the suffixed form is realized as [honor-is], and NOT as [honos-is]. Likewise, when we add the accusative suffix [-em], the whole result shows up as [honor-em] and NOT [honos-em]. Why? Well, once again, our /s/ is between two vowels and so it becomes an [r].

This whole scenario results in the following picture:

BEFORE THE s-> r rule AFTER THE s->r rule

honos 'nominative' honos 'nominative'

honos-is 'genetive' honor-is 'genetive'

honos-em 'accusative' honor-em 'accusative'

What's the big deal? The big deal is that after the "s-becomes-r-between-vowels rule" introduced its sound change in Latin, the Latin root meaning 'honor' is no longer the same throughout the paradigm. Sometimes it ends with an /s/, and other times it ends with an /r/.

At this point, the concept of paradigm leveling comes into play. Think of it this way. The sound /s/ in the nominative form [honos] looks around it and says, "hey, in all the other forms I'm an /r/. I'd better regularize myself here as well and become an /r/ also." This resulted in the following change:

AFTER PARADIGM LEVELING

honor 'nominative'

honor-is 'genetive'

honor-em 'accusative'

As you can see, on analogy with the other forms in the paradigm, the final sound in the nominative changed to /r/, even though it did not appear between vowels. It simply did so in order to make the root regular throughout the paradigm itself. That's paradigm leveling!

Interestingly, not all analogy results from regularizing pressures within the morphological system of a language. Some cases of analogy really don't have much to do with making the morphology more regular at all. Two of these that are kind of fun are back formation and folk etymology.

Back Formation

Back formation refers to a situation in which new base forms are created in a language. Basically, what happens is that speakers misanalyze the morphological structure of a word and wind up breaking the word apart to create a new word that didn't previously exist. A famous case in point is the verb "to burgle" in English--a verb which now is widely used and means roughly "to break and enter and steal". The word "burglar" was borrowed into English from from French after the Norman invasion of England. Think about how it sounds and you can understand what happened here. We have an -er suffix in English that means roughly, "one who VERBS", as in "runner-one who runs" or "painter-one who paints". Well, what happened with "burglar" was this. People erroneously analyzed burglar as comprising two morphemes: burgle + -er. Thus, they started using burgle as a verb, even though "burglar" was actually a borrowed word that was not made up of two morphemes.

So, what he have in the case of burglar is an example of analogy that works backwards! Rather than saying rhyme is to rhymed as climb is to X (climbed), people thought, "runner is to run as burglar is to X (burgle!!). That's why this process is called back formation.

Folk Etymology

Folk Etymology is also about misanalysis, but of a slightly different sort. In cases of folk etymology, people coin new terms in large measure because the morphology of the term that they are misanlyzing is not very clear to them. Your text book cites the example of a case at Ohio State where a writer used the term "garden snake" instead of "garter snake". What seemed to happen was this:

1) the term garter in the name held no clear relation (for this writer) to the snake in question

2) the term garter sounds a lot like garden

3) the writer associated this snake with gardens much more than garters

4) the writer formed a new word for this snake: garden snake

It may well be the case that the writer had "heard" garden snake all of his/her life, even though speakers may have been saying garter. Sometimes, the morphological roots of a word become a bit obscure and another morpheme seems to make equal or more sense and we fill the gaps. Of course, this mistake is an example of how folk etymology can work, but clearly it is not a case of a wholesale change in English. Last time I checked, we were still using the term garter snake. Think about how often you do this kind of thing when you think you understand the lyrics of a popular song, only to later see them printed on the inside of a CD.
 

Adding Words

One element of historical change involves the addition of words. Here are some of the processes that result in the addition of words to the lexicon.

Acronyms: NATO, scuba, radar (radio detection and ranging)

Blending: chortle (chuckle and snort)

Clipping: exam, lab, dorm, taxi cab (cabriolet)

Coinage: Kodax, Exxon, kleenex, Xerox, pooch and snob are all example of "coined" words

Conversion: think about taking the noun book and making it a verb. This is what historical linguists call "conversion". We saw this in morphology as a kind of derivational morphology regularly used in English.

Eponymy: naming things in terms of people connected in some way to them: hertz, watt, ohm, Washington, DC.

Borrowing: fiesta, siesta, beef, schlep,
 
 

Syntactic change

Besides phonology and morphology, historical change can also affect both the syntax and the semantics of a language. I'm going to highlight important areas of syntactic and semantic change in this section.

When we talk about syntactic change, what we're largely talking about is change in how words are ordered to form sentences in a language. In English, there has been a considerable amount of syntactic change if we look at the evolution of the language from the Old English period to the contemporary English that we speak now. As an example, in Old English, the possessive pronoun often followed the noun that it modified, so that "our father" could be said "father our" (abstracting away from sound changes, of course). Over time, however, the word order of English has become more fixed than it used to be. Now, we can't say "father our" unless we are trying to produce antiquated sounding English. In modern English, the possessive pronoun precedes the noun.

Your book points out that in the same way, English used to freely permit the "you" pronoun in a command to follow the verb. By contrast, nowadays, it would sound pretty funny if your parents said to you: "go you forward and find you a job!" These are examples of syntactic change in English. In particular, they are examples of a larger pattern in which English word order has become more restricted throughout the history of the language.

If any of you are familiar with German, a language closely related to English, you'll know that word order is MUCH freer. That is, speakers have much more liberty in moving the elements of a sentence around without fundamentally changing its meaning. The same is actually true of Spanish, as well. In Spanish, we can say "JUANA juega bien" or "Juega bien JUANA" and both sentences mean "Joan plays well." Compare that to English. Is it possible to say "Plays well JOAN" or are we limited to "JOAN plays well"?

One thing that people wonder a lot about is just what causes syntactic change. Looking for causes in such matters is actually a tricky business. One thing that seems to be true of English, though, is that as it lost its rich case system, it gained a more rigid word order. That is, as it lost morphological affixes that indicated the differences between subject and object and so forth, as well as affixes that marked agreement between subjects and verbs, its sentences gained a more inflexible ordering of elements. From a functional point of view, we might say that the stricter word order is a way of compensating for the loss of case. Now, we know a word is the subject because it comes before the verb in a declarative. To the extent that such observations are insightful, they also point to the interconnectedness of morphological and syntactic change.

Semantic change

As the name suggests, semantic change refers to changes in the meanings of words over time. There are a number of general ways that words change in meaning. Here's four types of change that are discussed in Language Files:

• semantic extensions
• semantic reductions
• semantic elevations
• semantic degradations

Semantic extension

When we say that a word's meaning has been extended, what we mean is that the range of uses--i.e. the things that it refers to and the contexts in which its use is appropriate--has grown over time. A nice example is the case of Western Apache (not discussed in your textbook). Western Apache was obviously around before the automobile. So, when the automobile arrived on the scene, it's pretty evident that Western Apache wasn't going to have a bunch of words sitting around to refer to auto parts such as headlights, windshield, and so forth. What did they do? They simply extended the meanings of their words for body parts. So, the same word that is used for "eyes" is also the word for headlights. This is a nice case of extension, because it is clear that the use of the Western Apache word for "eye" was extended beyond its reference to the eye of an animal.

Extensions happen ALL THE TIME. Your book cites the example of "nuke" which has been extended from something like a verb meaning "to drop an atomic bomb on" to a broader, less specific meaning of "destroy". In fact, "nuke" has been even more extended. For me, at least, "nuke" is also a verb meaning "to cook or heat in a microwave oven." How about for you? Words like Kleenex and Xerox are examples of brand names that got extended to mean more generally "facial tissue" and "photocopy", respectively. Although, interestingly, I think that Xerox is not as widely used as it used to be. More often than not, I here people use the term "photocopy" rather than Xerox these days.

One thing about extensions is that they are often metaphorical. That is, the extended meaning is conceptually similar to the original meaning. Your book gives the nice example of "broadcast". Originally, this word referred to scattering seed over a field. Now, it refers to transmitting radio waves through space. The comparison is pretty clear. Interestingly, in this case, the original meaning has been lost, so the extension feels now simply like a full fledged shift in meaning.

Semantic reductions

This is the opposite of extension. Rather than gaining a broader range of meanings, words that under reduction gain more specific meanings. In Old English, [hund] referred to dogs in general. Today, the word 'hound' refers to a particular type of dog--a subset of the larger set of dogs. Its meaning has been reduced.

Semantic elevations

Sometimes, words gain prestige through time. The word "knight" is a nice example. In Old English, the word meant "youth" or "military follower". Now, if you live in England and are knighted, you join the ranks of the nobility. Clearly, there's more prestige in being a Knight today than there was then.

Semantic degradations

If some words gain in prestige, others take on more negative connotations. The word "silly" in English historically meant something like "happy, innocent", but now it means something more like "dumb, foolish". So, here's a case of a word that took on more negative meanings through time and lost its original, more positive meaning.

Though we can do a pretty good job of tracking the nature of change in meaning, the question of why meanings change is somewhat murkier. First of all, though we know that meanings change over time, we really can't predict whether a given word will change or how it will change. There was no way to know, for example, that the English word "girl", which once referred to children of both sexes, would reduce its meaning such that now it refers only to female children.

Additionally, it is not clear that there is a single answer to the question of "why" words change. In fact, I think it is more useful to turn the question on its head and ask: "why shouldn't meanings change?" I'd be very surprised if meaning (as well as phonology, syntax, and morphology) didn't change over time. Remember that words are not the things that they refer to. The whole notion of just what meaning is is a fairly slippery area. People use words creatively to express novel thoughts and ideas, to draw comparisons, etc... The meanings of words do not exist independently of the people who use them. If words are viewed as tools for communication, then we might expect to see that the tools are freely changeable by the users in order to better serve their communicative needs. Essentially, the meanings of words and the order in which we use them are negotiated by us all as a speech community. This is true in all languages. And there are any number of factors--both grammatical and sociological--that can come to bear on why meanings might change.

While we're on this point, meaning change is one of the areas that so-called language experts (the people who supposedly tell us how badly we all speak and who tell us that English is rapidly going to hell in a handbasket) get most exercised about. Let's take a fairly simple example. I've heard a bunch of people rail against the use of "hopefully" in sentences like this:

Hopefully, I'll get an A on my Linguistics exam next week.

It's not really that they don't want you to get an A on the exam. What they get excited about is the use of the adverb "hopefully." It's "supposed" to mean something like "done with hope", they say, but in this sentence, the speaker is using it to mean "I hope that". For them, "hopefully" should be used like this:

Hopefully, I went to see my grandmother in the hospital.

Sorry, but I use "hopefully" the way most of us do. I say things like: "Hopefully, the Jets will make the playoffs." or "Hopefully, we'll get some rain before the month is over," or "Hopefully, coach K will have to watch UNC win the NCAA tournament this year." What's going on here? Simple. "Hopefully" has undergone a shift in meaning (in my dialect of American English, at least). I'm not wrong to use it that way. I'm using in in perfect accordance with the meaning that "hopefully" has in my grammar--a meaning that is clearly shared by most of you, given out poll in class. Finally, the hopefully case makes one more point. Not all meaning change necessarily involves "elevation" or "degradation". Sometimes it's even difficult to say whether a meaning shift is either an extension or a reduction. In the case of "hopefully", there seems to be something more akin to a subtle shift that is going on, away from one use and towards a similar though still distinct use of the word.