Historical Linguistics 3

What is the primary criterion for subgrouping languages?

The primary criterion is shared innovations—unique changes that cannot plausibly have occurred independently in unrelated languages, and must be unusual enough to rule out convergent evolution or chance.

Provide an example of a shared retention.

Latin *pater* and Sanskrit *pitā* both retain the Proto-Indo-European form for "father," but this does not prove a special relationship between Latin and Sanskrit beyond their shared ancestry.

Provide an example of a shared innovation.

Grimm’s Law sound changes, such as Proto-Indo-European *p becoming Germanic *f, are unique to Germanic languages and prove they form a subgroup within Indo-European.

Why are single phonetically natural sound changes poor evidence for subgrouping?

Because such changes frequently occur independently in unrelated languages, like the change from [w] to [v] in Latin, Norse, and High German at different times.

Why are ordered sequences of sound changes powerful tools for subgrouping?

The probability of the same sequence occurring independently in unrelated languages is extremely low; for Germanic languages, the probability of all 7 ordered changes occurring by chance is about 1 in 266,667, confirming a valid subgroup.

What are the 7 ordered sound changes that define the Germanic subgroup?

Grimm’s Law parts 1, 2, and 3, Verner’s Law, fricatives becoming stops after nasals, stress shift to the initial syllable, and unstressed *e changing to *i unless followed by *r.

Why is the mathematical rarity of the Germanic sound changes significant?

The probability of all 7 changes occurring in the same order by chance is approximately 1 in 266,667, confirming that Germanic languages share a common history and form a valid subgroup within Indo-European.

How do morphological innovations contribute to subgrouping?

Morphological innovations are highly valued because they are often idiosyncratic, such as the Germanic weak past tense and the double paradigm of adjectives in Germanic languages.

Why are syntactic rules the least useful for subgrouping?

Syntactic rules are less useful because there are limited ways to structure grammar, similar patterns can develop independently in unrelated languages, and syntactic changes are less regular and harder to trace.

What are the two models of language diversification?

The Tree Model, best for clean diversification where dialects lose contact quickly, and the Network Model, necessary when dialects remain in contact and trade innovations in overlapping patterns.

How can a network model appear as a tree model?

If the intermediate linking dialects go extinct, leaving only distinct survivors, a network can appear as a clean tree, such as if dialects A, B, and C form a network but B goes extinct, making A and C appear as sister languages.

How do modern computational tools assist in subgrouping?

Computers can evaluate thousands of linguistic traits to see how well they fit various possible trees, account for uncertainty, and investigate relationships beyond traditional manual methods.

What are proto-lexica?

Proto-lexica are reconstructed vocabularies of protolanguages, such as Proto-Indo-European or Proto-Algonkian, used to infer cultural, ecological, and linguistic relationships among ancient speech communities.

What is the optimistic goal of proto-lexical research?

To demarcate the geographical area where a protolanguage was spoken and connect findings to archaeology, such as material culture and climate, by reconstructing enough terms for plants, animals, and weather.

What are the four challenges in using proto-lexica for cultural or ecological inference?

False reconstructions, gaps in the lexicon, ignoring cladistic structure, and semantic change.

What is a false reconstruction in proto-lexica?

A reconstructed word that appears valid but is misleading because it was not actually present in the protolanguage, such as Proto-Algonkian *paaškesikani*, which was coined in one daughter language and translated morpheme-by-morpheme into others.

Why are false reconstructions a problem for proto-lexical inference?

They can mislead researchers into making incorrect cultural or ecological inferences, such as assuming a protolanguage had a word for a concept it did not.

What is the problem of gaps in the lexicon?

Some basic words are missing in proto-lexica even though they must have existed in the protolanguage, such as no Proto-Indo-European word for "iron" or "finger."

What does a gap in the lexicon imply?

A gap does not imply cultural absence; for example, the lack of a Proto-Indo-European word for "finger" does not mean PIE speakers had no fingers, but that the word was lost or replaced over time.

Why is lexical replacement common over time?

Due to semantic shift, borrowing, analogy, and taboo, such as the word "silly" originally meaning "blessed" but now meaning "foolish."

What is the problem of ignoring cladistic structure in proto-lexical studies?

It can lead to incorrect inferences, such as Siebert’s Proto-Algonkian Homeland Study, which did not include Blackfoot, leading to conclusions that apply only to non-Blackfoot Algonkian.

Why is cladistic structure important in proto-lexical inference?

Because not all daughter languages share the same history, and ignoring the family tree can lead to incorrect reconstructions of the protolanguage’s geographical or cultural context.

What is the problem of semantic change in proto-lexical inference?

Words change meaning over time, making their original referents uncertain, such as the Proto-Indo-European word for "beech" possibly referring to oaks or any edible-nut tree.

How can semantic change distort proto-lexical inference?

Through meaning shifts, metaphor or metonymy extensions, and borrowing, all of which can alter a word’s meaning over time.

What is the goal of long-distance language comparison?

To reconstruct relationships beyond well-known families, such as Proto-Indo-Uralic or Nostratic, using proto-lexica.

What are shared features proposed for the Indo-Uralic hypothesis?

Pronouns like first person *m- and second person *t-, noun endings like the accusative *-m, and basic vocabulary such as Nostratic *nimi ("name") and *weti ("water").

What is the main problem with long-distance language comparison?

Lexical replacement over time erodes evidence, making it difficult to prove distant relationships.

What is the problem of chance resemblances in long-distance comparison?

Chance matches are inevitable in large wordlists, leading to false positives, such as Spanish *mucho* and English *much*, which look and mean the same but are not historically related.

Provide three examples of chance resemblances.

Spanish *haber* and English *have* are not related. Spanish *día* and English *day* are not related. Spanish *bola* and English *ball* are not related.

What is the lexical retention problem?

Languages lose inherited words over time, making it harder to prove distant relationships, such as English and German sharing 82 cognates, while French and Tocharian B share only 27 cognates.

What is the implication of the lexical retention problem?

Beyond approximately 5,000–8,000 years, shared vocabulary is roughly equal to chance resemblances, with the limit of the comparative method being about 10,000 years; older relationships cannot be proven with current methods.

What are the three responses to the challenges of long-distance comparison?

Older Nostraticists claim chance resemblances are rare, but this is false. Mass comparison compares large wordlists for patterns but lacks regular sound correspondences. Statistical methods assume a constant rate of lexical replacement, which is not true.

What is the barrier of random change in long-distance comparison?

Over time, languages change randomly, and inherited vocabulary shrinks, so eventually shared words become indistinguishable from coincidence, and relatedness can no longer be proven.

What is convergent evolution in linguistics?

The independent development of similar features in unrelated languages due to similar environmental or social pressures, such as Spanish *mucho* and English *much*, which look and mean the same but are not historically related.

What is the hypergeometric distribution in linguistics?

A rigorous mathematical tool used to determine whether similarities between two languages are due to shared ancestry or random chance, serving as a foundational component of modern statistical methods for long-distance language comparison.

Provide the example of hypergeometric testing in Tocharian B and English.

In a wordlist of 108 meanings, English has 10 words starting with /f/ and Tocharian B has 12 starting with /p/, with 6 matches such as *father* and *pācer*, and a probability of less than 1/100,000, concluding a genuine genetic relationship.

What are the 6 English-Tocharian B cognates with initial /f/ and /p/?

Father and pācer, feather and paruwa, fire and puwar, five and piś, fly and plu-, foot and paiyye.

Why does relaxing rules weaken hypergeometric testing?

Because it doubles or more the probability of a chance match, such as allowing /f/ to match /p/, /b/, or /v/, making it harder to cross the significance threshold.

Why is the Monte Carlo method useful for long-distance comparison?

Because it accounts for chance resemblances by generating a null distribution, provides a visual representation of random vs. real similarities, and helps distinguish between genetic relationships and coincidence.

What is the individual-identifying threshold in linguistics?

A statistical threshold for determining whether a linguistic trait is too rare to occur by chance, with approximately 7,000 human languages, a trait with a 1-in-700,000 chance being virtually impossible to appear in two languages by accident.

Why are grammatical phenomena better indicators of a common ancestor than lexical items?

Because lexical items can be borrowed, while grammatical phenomena, such as prefix paradigms or word order rules, are too complex to arise by chance or borrowing.

What is the conclusion of the Algonkian Prefix Paradigm study?

The extreme rarity of the Algonkian prefix paradigm proves that it is a shared innovation from a common ancestor, not a random similarity.

What are three types of change?

The three types of change are phonological change, syntactic change, and morphological change.

What is a phonological change?

Phonological change refers to changes related to the sound system of a language.

What is syntactic change?

Syntactic change refers to changes related to the structure and rules of sentence formation in a language.

What is morphological change?

Morphological change refers to changes related to the forms and inflections of words, such as the simplification of plural endings or the consolidation of verb forms.

What is NLA?

NLA (Native Language Acquisition) is the process by which children learn their first language by constructing systems of rules. Most changes in linguistic structure begin as errors made by native learners during this process.

What did Chomsky (1970) observe about English deverbal nouns and gerunds?

Chomsky noted that gerunds (e.g., *destroying*) are formed by a single productive rule, while deverbal nouns (e.g., *destruction*) are lexically idiosyncratic, suggesting derivational morphology requires lexicon-based rules.

Why was the lexicon enriched in generative morphology?

To accommodate idiosyncratic word formation, such as deverbal nouns, which do not follow uniform rules like gerunds.

What is lexical phonology, as proposed by Kiparsky?

A theory where morphological and phonological rules interact cyclically within the lexicon.

Where did Halle (1973) place inflectional morphology?

In the lexicon, though this raised questions about interactions between syntax and the lexicon.

What are the alternative views on the placement of inflectional morphology?

Some argue it belongs in syntax (e.g., Spencer 1991) or phonology (e.g., Anderson 1982).

What is an example of a morpheme with no definable meaning?

The English morpheme *-ceive* (e.g., *conceive, deceive*), which has no independent meaning but is a formal unit due to its allomorphy.

What is the "peg prefix" in Navajo, and what is its function?

The morpheme *yi-*, which is semantically empty but serves a phonological function in verb prefixes.

What is the "head" in derivational morphology?

The component of a derived word that determines its lexical class and interacts with syntactic rules, often the last category-changing affix added.

What does Baker’s Mirror Principle state?

The order of morphological affixation mirrors the order of syntactic operations, supporting the idea that word formation is word-internal syntax.

What is Distributed Morphology, and why is it considered promising?

A theory that treats most morphology as syntax while acknowledging an autonomous module for non-syntactic phenomena, integrating morphology, syntax, and phonology into a cohesive framework.

What does it mean that the lexicon in DM is modularised?

In DM, the lexicon is divided into two main components: the Encyclopedia, which stores idiomatic or irregular forms, and the Vocabulary, which contains phonological strings paired with morphosyntactic features for insertion at syntactic terminal nodes.

What are phonological strings?

Phonological strings are the actual pronunciations paired with morphosyntactic features in the Vocabulary. They are inserted at terminal nodes after syntax is complete.

What are morphosyntactic features?

Morphosyntactic features are grammatical properties, such as tense, number, or case, that are specified at syntactic terminal nodes and later filled with phonological material from the Vocabulary.

What are syntactic terminal nodes?

Syntactic terminal nodes are the end points of syntactic trees where morphosyntactic features (e.g., [+plural], [+past]) are specified. These nodes are later filled with phonological material from the Vocabulary.

What are root nodes?

Root nodes are morphemes without inherent category features (e.g., *dog*, *run*). They acquire category (noun, verb) through syntactic context or other morphemes (e.g., *-er* in *runner*).

What are major word classes?

Major word classes are categorizing heads, such as nouns, verbs, and adjectives, that determine the syntactic category of a word. Examples include *-ness* (noun-forming) or *-able* (adjective-forming).

What are intermediate cases?

Intermediate cases are morphemes that blend properties of both inflection and derivation, such as participles like *broken*, which can function as adjectives or verbs.

What does it mean that DM does not distinguish between inflection and derivation?

DM does not recognize a strict distinction between inflection and derivation because intermediate cases, such as participles and gerunds, are common, making it impractical to draw a clear line.

What structure do words have in DM?

In DM, words have a double structure: a hierarchical syntactic tree and a string of phonological forms inserted at terminal nodes. Phonological rules can impose hierarchical structure, but it is not directly derived from syntax.

What are empty nodes?

Empty nodes are nodes inserted for morphological well-formedness but with no phonological content. For example, in *sheep* (singular/plural), the plural node may be empty, with no overt *-s*.

What is syntactic output?

Syntactic output is the hierarchical structure generated by syntax before Vocabulary insertion. It includes morphosyntactic features, empty nodes, and fused nodes.

What is the role of syntactic structure in DM?

Syntactic structure, including morphosyntactic categories, is autonomous and abstract. No syntactic structure is projected from lexemes, and phonological material is inserted after the syntactic tree is formed.

Why does DM avoid projecting syntactic structure from lexemes?

Because word-internal structures follow fixed patterns rather than idiosyncratic meanings, and generating them with the same machinery as external syntax is more economical.

What are the two main components of the lexicon in DM?

The two main components of the lexicon in DM are the Encyclopedia, which stores meaningful units or idioms, and the Vocabulary, which contains phonological strings paired with morphosyntactic features for insertion at syntactic terminal nodes.

What are f-morphemes and l-morphemes in DM?

F-morphemes (functional morphemes) have no choice of insertion (e.g., inflectional markers like *-s* for plural), while l-morphemes (lexical morphemes) allow multiple insertion options (e.g., roots or major word classes like *dog* or *run*).

How does DM handle the distinction between inflection and derivation?

DM does not recognize a strict distinction between inflection and derivation, as intermediate cases (e.g., participles, gerunds) are common, making it impractical to draw a clear line.

What is the double structure of a word in DM?

A word in DM has a hierarchical syntactic tree and a string of phonological forms inserted at terminal nodes. Phonological rules can impose hierarchical structure, but it is not directly derived from syntax.

What processes can occur before Vocabulary insertion in DM?

Before Vocabulary insertion in DM, morphosyntactic features can be added, nodes can fuse, empty nodes can be added for morphological well-formedness, or feature impoverishment can occur.

What are readjustment rules?

Readjustment rules are post-syntactic phonological operations that alter already inserted vocabulary items in a specific morphosyntactic context, such as past tense or plural forms.

What is process morphology?

Process morphology refers to phonological or morphological operations, such as umlaut or ablaut, that apply after Vocabulary insertion to adjust forms, like *mouse* to *mice* via i-umlaut.

What is suppletion?

Suppletion is the use of a word as a particular form of a verb when the word is not related to the main form of the verb. For example, *went* is used as the past tense of *go*.

What is DM's stance on paradigms?

DM treats paradigms as epiphenomena, or side effects, of syntactic and phonological rules. Paradigms emerge from the interaction of syntactic feature specifications, Vocabulary insertion, and post-syntactic readjustment rules.

What is resegmentation?

Resegmentation occurs when speakers or learners redraw the boundaries between parts of a word. A form that earlier speakers treated as one unit may later be interpreted as two units, or a boundary may shift.

What does splitting terminal nodes mean?

Splitting terminal nodes means a single terminal node is split into two, assigning part of the original phonological string and new content to each, such as *movable* being split into *move* + *-able*.

What does backformation mean?

Backformation is the process of creating a simpler word from a more complex one, such as forming *donate* from *donation* or *orientate* from *orientation*.

What is univerbation?

Univerbation occurs when separate words become a single word, such as the German word *aufführen*, which comes from *auf* + *führen*.

What was i-umlaut before it became a grammar rule?

Before it became a grammar rule, i-umlaut was a physical habit of speech in Old English.

What is i-umlaut?

I-umlaut is a single sound-change rule in Old English that created vowel mutations across almost all types of words, such as *mūs* (mouse) to *mŷs* (mice).

What is the Shrinking Rule?

The Shrinking Rule refers to the process in Middle English where speakers began stripping the i-umlaut rule away to make the language simpler.

What is Verb Cleanup?

Verb Cleanup is the process where the i-umlaut rule was completely erased from action verbs in Middle English. Vowel changes, vowel shortening, and blending consonants together were thrown out to make the language more uniform.

What is an inflectional paradigm?

An inflectional paradigm is a complete set of inflected forms of a single lexeme, showing how a word changes to express different grammatical categories like tense, number, or case.

What is syncretism in historical linguistics?

Syncretism in historical linguistics is the phenomenon where distinct morphosyntactic categories are expressed by the same phonological form.

What are three main origins of syncretism?

The three main origins of syncretism are phonemic merger, learner errors, and a combination of both.

What does the Impoverishment Filter rule state?

The Impoverishment Filter rule states that it is a rule that suppresses a morphosyntactic feature, causing the default (unmarked) form to surface.

What does default/unmarked form refer to?

The default or unmarked form refers to the most basic or neutral form of a word or grammatical category that is used when more specific forms are not required or have been lost.

What does syncretism by default mean?

Syncretism by default means that the default or unmarked form replaces functionally redundant variants.

Explain syncretism in Gothic passive.

In Gothic passive, syncretism is evident because all non-singular forms use the same ending, specifically the 3rd person plural ending, which serves as the default form for all non-singular contexts.

What does non-singular syncretism mean?

Non-singular syncretism means that all non-singular forms use the same grammatical ending or form, often the default form, instead of having distinct forms for each person or number.

Explain syncretism in West Germanic.

In West Germanic, syncretism occurred where number outranked person, so losing person marking caused a default to the 3rd person plural form rather than the 1st person singular.

What does marking mean in the context of syncretism?

Marking in the context of syncretism refers to the presence of distinct morphological features that indicate specific grammatical categories.

What does it mean that a paradigm is active?

An active paradigm means that the set of inflectional forms for a lexeme is productively used and can undergo changes, such as syncretism or leveling.

What is leveling?

Leveling is the elimination of an alternant occurring in a specific morphosyntactic environment, where a default marker replaces a functionally redundant variant.

Explain gender syncretism.

Gender syncretism is the process where distinct gender forms merge into a single form.

What is suppletion?

Suppletion is the use of a word as a particular form of a verb when the word is not related to the main form of the verb.

What is a paradigm in the context of historical linguistics?

A paradigm in historical linguistics is a complete set of inflectional forms of a single lexeme.

What are suppletives?

Suppletives are synonymous defective lexemes that are never in functional competition.

What is a concord class?

A concord class is a grammatical category that divides the nouns of a language into arbitrary inflectional classes for the purpose of syntactic agreement.

What are lexeme classes (inflectional classes)?

Lexeme classes, or inflectional classes, are sets of lexemes whose members follow the same pattern of inflectional forms.

What is a default class?

A default class is the most common or unmarked inflectional class in a language.

What is a class in linguistics?

A class in linguistics refers to a group of items, such as words or morphemes, that share the same morphological, syntactic, or phonological behavior.

What is a Sound Change?

A sound change refers to spontaneous changes in pronunciation that occur over time in a language.

How do sound changes spread?

Sound changes spread through acquisition error, variable change, social marker, and categorical rule.

Define acquisition error.

Acquisition error is a spontaneous misperception or articulatory failure by a native language learner.

Define variable change.

Variable change is when the error is copied and survives as a variable sound change among native speakers.

Define social marker.

A social marker is a linguistic feature that is copied and survives as a variable sound change among native speakers, often indicating social identity or group membership.

Define categorical rule.

A categorical rule is when learners acquire the statistical pattern of a sound change, completing the shift without lexical exceptions.

Where do sound changes come from?

Sound changes come from failure to distinguish sounds, mishearing speech errors, easier pronunciation, and mishearing acoustic signals.

Give an example of failure to distinguish sounds (merger).

An example is when a child stops hearing two sounds as different, such as merging the sounds in *cot* and *caught*.

Give an example of mishearing speech errors.

An example is when people make pronunciation shortcuts, and children may think those shortcuts are correct, such as saying *wanna* for *want to*.

Give an example of easier pronunciation (articulatory ease).

An example is when some sound combinations are hard to pronounce, so children may not learn to produce them, such as pronouncing *nuclear* as *nucular*.

Give an example of mishearing acoustic signals.

An example is when listeners may confuse sounds because they sound similar, such as confusing *ask* with *ax* in some dialects.

What are characteristics of sound change?

Shifts in pronunciation are gradual, subject to fine phonetic conditioning, and exhibit no lexical irregularities.

What is lexical diffusion?

Lexical diffusion is the gradual spread of a sound change across the lexicon, affecting words at different rates rather than uniformly.

What is the Great Vowel Shift?

The Great Vowel Shift is a major change in English pronunciation that occurred mainly between the 14th and 16th centuries.

When did the Great Vowel Shift happen?

The Great Vowel Shift happened mainly between the 14th and 16th centuries.

What changes happened during the Great Vowel Shift?

During the Great Vowel Shift, mid vowels rose, and the long high vowels were diphthongized.

What is lenition?

Lenition is a common conditioned sound change that typically affects consonants between vowels, resulting in less articulatory interruption.

What is the "ease of articulation" theory?

The "ease of articulation" theory suggests that sound change was often attributed to speakers aiming for less articulatory effort.

What is Articulatory Phonology?

Articulatory Phonology is a theory that views speech as a series of overlapping gestures, with sound changes arising from mistakes in gesture timing.

What is slippage in timing?

Slippage in timing refers to errors in the overlap of articulatory gestures, such as aspiration occurring if the glottis stays open too long after a mouth closure.

What is target undershoot and overshoot?

Target undershoot is when a speaker fails to reach an articulatory target, and target overshoot is when a gesture is exaggerated.

What is the Old Rule (Rhotacism)?

The Old Rule, or Rhotacism, is a sound change that changes the sound "s" into "r" if it was trapped between two vowels.

What is the New Rule (Simplifying "ss")?

The New Rule is a sound change where "ss" turns into "s" if they came after a long vowel or a diphthong.

What are the types of sound change?

The types of sound change are unconditioned change, conditioned change, merger, loss, and split.

Define unconditioned change.

Unconditioned change is when a sound changes everywhere, regardless of the sounds around it.

Define conditioned change.

Conditioned change is when a sound only changes in specific environments.

Define merger (sound change).

Merger in sound change is when two different sounds become the same sound.

Define loss (sound change).

Loss in sound change is when a sound simply disappears.

Define split (sound change).

Split in sound change is when one sound becomes two different sounds.

What does the Invariant Transparency Hypothesis say?

The Invariant Transparency Hypothesis states that children learn the sounds they hear as the underlying forms of words without inventing complex rules.

Name potential sources of language change.

Potential sources of language change are language contact, deliberate manipulation, and transmission from generation to generation.

Why is the source of any particular change elusive?

The source of any particular change is elusive because historical records are often impoverished, and most changes begin as idiosyncrasies in individual speech.

What is NLA?

NLA stands for Native Language Acquisition, which is the process by which children acquire their first language.

What is the Critical Period Hypothesis?

The Critical Period Hypothesis is the hypothesis that a native language must be acquired within a developmentally defined "critical period."

What is the cycle of language replication?

The cycle of language replication is the process by which a language continues to exist through native language acquisition across generations.

How are major changes in phonology, syntax, and morphosyntax brought about?

Major changes in phonology, syntax, and morphosyntax are brought about as learner errors, not conscious borrowings.

Why do major changes in phonology, syntax, and morphosyntax happen because of learner errors?

Adults can easily borrow vocabulary and expressions, but they struggle to fully borrow phonological contrasts and grammatical structures.

What is subgrouping in historical linguistics?

Subgrouping identifies lines of descent in a language family, determining which languages remained unified after initial diversification, with subgroups as nodes in a tree diagram.

Why are shared retentions not useful for subgrouping?

Shared retentions are inherited traits any daughter language could preserve, proving only a common ancestor, not divergence from an intermediate one.

What is the mechanics of the hypergeometric distribution?

It calculates the probability of observed correspondences between wordlists using N (total meanings), n (sound a in A), R (sound b in B), and r (tokens of correspondence).

What is the paradox of list length in hypergeometric testing?

Adding less basic words increases noise; shortening lists makes it harder to cross significance thresholds.

What is the Monte Carlo method in linguistics?

A method to prove language relations by comparing wordlists, calculating similarity scores, and creating null distributions through repeated randomization.

What is the Algonkian Prefix Paradigm case study?

Algonkian uses four prefixes (*ne-*, *ke-*, *we-*, *me-*). The probability of two languages independently choosing this sequence is ~1-in-1,000,000.

What is the conclusion of the Algonkian Prefix Paradigm study?

The rarity of the Algonkian prefix paradigm proves it is a shared innovation from a common ancestor.

What does analogy refer to in linguistics?

Analogy is when one word form influences another to make language more consistent, e.g., "preached" leading to "beseeched."

How does DM redefine the concept of a "word"?

DM posits words have hierarchical structures generated by syntax, with boundaries as superficial constraints, not theoretical entities.

How are morphemes classified in DM?

Morphemes are f-morphemes (functional, no insertion choice) and l-morphemes (lexical, multiple insertion options).

What is reinterpretation?

Reinterpretation is when speakers reanalyze word parts, creating new morphemes, e.g., Old English *dom* becoming the suffix *-dom* in *boredom*.

Explain affixes becoming clitics.

Affixes evolve into clitics attaching to phrases, e.g., English *'s* in *John's book* vs. *the king of England's crown*.