Professional Documents
Culture Documents
Defining Morphology: Learning Objectives
Defining Morphology: Learning Objectives
Defining Morphology
AL
RI
Learning Objectives
TE
MA
Define what morphology is as a field.
Tackle the challenges of defining a “word.”
Learn the basics of the three-line gloss.
D
0. Preliminaries
Morphology, at its core, is the study of the word. Perhaps outside of a linguistics class, we
RI
rarely grapple with the question of “what is a word?” In a game of Scrabble or in another
context where someone uses a term we are unfamiliar with, we may turn to one of our
PY
favorite dictionaries to confirm something’s wordhood. Even then, we recognize that the
dictionary is slow to adapt to new usages and coinages – further, it does not always rec-
ognize what is basically productive about our linguistic system. For instance, if your boss
CO
instructed you to “recollate the files in the conference room,” you might grumble to yourself
about the tediousness of the task. However, you probably would not point out that
Merriam-Webster Dictionary does not recognize “recollate” as a word of English –
therefore the task makes no sense. You certainly would not point this out if you planned
on keeping that job for any time.
Wordhood is not defined by whether or not the given item is listed in an official
ocument. In our recollate case, we recognize that this is an English word, even if it is
d
completely novel to us, because it follows predictable rules of English word formation:
we combine two separated elements re-(a prefix meaning “do again”) with collate (a verb
stem meaning “to put into order”). This prefix re-generally may occur with verbs and as
speakers of English, we are aware of this and may use this to produce new and novel
forms. Thus, our mental capacity for words may generate new forms and may do so far
faster than any physical dictionary could keep up with. This mental capacity is part of
our mental grammar. It is this grammar that we are studying throughout this book.
1. Wordhood
Much as we must discard the idea that wordhood is associated with being listed in an
official registry (a dictionary), we must consider and reject a few other common ideas
about defining a word. In the first place, as readers of English, one common assumption
about wordhood is that a word is a unit of language, which has spaces on both sides of it.
This may be an adequate description of the general situation with English writing (though
later in the book we will argue that complex terms with many spaces like pickle jar lid
factory shift supervisor are, in fact, a single word). However, it certainly fails to account for
spoken English, which does not put a pause between each word, let alone languages with
other written traditions or the roughly 70% of languages that linguists estimate have no
major written tradition.
Fun Fact: While European writing traditions now involve spacing between words, this
was not always the case. Early medieval and classical writing was done in a tradition
known as scriptio continua. In this style of text, breaks were not found between words but
instead where pauses would naturally be found, even within a word, when the text was
being spoken.
(1) blackbird
But, perhaps even more compelling, we may look to other languages of the world where
a single word may express the equivalent of an English sentence. Such languages are
known as polysynthetic, and we will look more at their properties throughout this book.
Chapter 1: Defining Morphology 5
An illustrative example is given below from the language Central Alaskan Yup’ik
provided by first-language speaker George Charles via Mithun (1999):
(2) kaipiallrulliniuk
kaig -piar -llru-llni -u -k
be.hungry -really -pst -apparently -indicative -they.two
“the two of them were apparently really hungry”
To be able to read and interpret all of the details of this example is not yet critical for us,
though we will soon begin to cover the concept of glossing. What is critical about this
example is that what would be a full sentence made up of independent words in other
languages, such as English, is a single word in Central Alaskan Yup’ik. Further, this pro-
cess of word building is fully productive in languages like Central Alaskan Yup’ik.
We will use the term productive throughout this text although there is not a precise
definition for what it means. Generally, the term is used to describe morphological pro-
cesses that are common in the language or morphological forms that occur regularly. We
contrast it with nonproductive that describes morphological processes and forms that are
found in a language but only rarely – perhaps due to historical change or language con-
tact. For example, the English plural marker -s is productive because it is commonly
found and applies to new words introduced to the language. The English plural marker
-ren found in children is nonproductive as it is limited to one exceptional form.
What, then, makes a good definition for “word?” Haspelmath (2011) makes a com-
pelling case that it may be impossible to generate an acceptable definition that will
account for all languages. Nonetheless, we will continue to use this term “word” in an
informal sense, but we will need to exercise caution making any formal reference to the
concept of a word. What we should notice is that regardless of the status of the concept of
“word,” languages build meaningful units out of other smaller meaningful units.
Sometimes those units may stand on their own and be meaningful or usable as in the case
of black and bird in blackbird, but in many cases, these units must combine with other units
to be usable such as re-in recollate. These units, termed morphemes, are ultimately what
this book is about. In many introductory textbooks, morphemes are defined as something
like the “smallest unit of meaning.” Throughout this book, we will question and refine
this definition. But, in a general sense (for now), a morpheme is the smallest of the units
of meaning in language. An independent word containing no sub-morphemes, such as
bird, is itself a morpheme. Our goal is to understand the universal properties, distribu-
tions, and restrictions of morphemes. We will also study how they interact with other
systems of grammar, such as phonology and syntax.
The study of morphology goes well beyond defining the concept of words. Our study
will involve analyzing the internal components of words and how these components
interact with other grammatical systems. It is our hope that by understanding morphol-
ogy and its role in our broader linguistic systems we can better understand the nature of
language itself.
6 Understanding Morphology
2. Doing Morphology
Linguists study morphology by breaking apart language (words, phrases, and sentences)
into smaller component pieces: morphemes. Ultimately, this work is a form of scientific
hypothesis building and is subject to the same scientific process as any other science.
Scientific method
Observe
Hypothesize
Predict
Test
A critical point is that just as there is no master list of words, there is no master list of
morphemes. Doing morphology is a process of scientific discovery, which includes trial
and error. Not every hypothesis will be correct. Sometimes, elements will appear to be
morphemes when they are not or will appear to not be when they are. It is only through
repeated observation and testing of our hypotheses that can we begin to be sure.
Morphology, like all scientific practice, is governed by the principle of simplicity of
explanation, also known as Occam’s Razor. The basic idea is that if you have multiple
viable competing solutions to the same problem, whichever requires the least amount of
explanatory effort is most likely correct. In morphology, this typically means that we look
to have fewer rules and constraints and we like to avoid having the same morpheme serving
different functions without a clear understanding of why.
Occam’s Razor: Occam’s Razor is named after the English medieval theologian
and philosopher William of Ockham (Occam) to whom the idea is most often
attributed. Whether he ever overtly expressed these ideas is questionable (see
Thorburn 1918). But the core idea remains a guiding principle of both scientific and
philosophical thinking.
Chapter 1: Defining Morphology 7
Interlinear Glossing
The first line of an interlinear or three-line gloss is in the target language, which is the
language that is being studied. The target language is typically written in IPA, the lan-
guage’s orthography if it is sufficiently close to the metalanguage’s, or an orthographic
representation like that of the metalanguage. This line is annotated with special symbols
to mark where the linguist believes morpheme boundaries exist.
Sometimes there is a line provided above this line, which provides the target lan-
guage’s orthography or an example without the morpheme boundaries. We saw an
example of this optional fourth line in our Central Alaskan Yup’ik example in (2). Again,
the primary goal in glossing is to be maximally informative to your potential readers
about your hypotheses about the structure of the language’s structure.
The second line of an interlinear gloss is the item-by-item translation. This line is
written in the metalanguage, which is a term for the language of analysis, along with the
conventions of glossing. There are numerous conventions of this line and the interaction
of the first and second line is the most critical to understanding an interlinear gloss.
Line three is the free translation. This line is also written in the metalanguage, but it
is a looser translation. It is easiest to think of this as the type of translation you might find
in a guidebook or other similar travel guide.
We will spend much more time on glossing during the next few chapters, but let us exam-
ine the basics of a gloss. Recall that the goal of a gloss is to hypothesize about the morpho-
logical structure of a given word or phrase. Even though we questioned the s tatus of the
word to begin this chapter, we will continue to utilize the concept in our glosses. Each
word boundary is marked by a space. Morphemes found internal to words are separated
with a hyphen.
A period is used in glossing when there are multiple functions or concepts contained
within a single morpheme. This concept is illustrated in the data below from Hiaki
(Uto-Aztecan, Arizona, and Sonora) taken from Bobaljik and Harley (2017).
Again, recall that the left edge of each Note that while we must be
word is aligned between lines 1 and 2 sure to make sure there are
matching numbers of hyphens
between corresponding words
(4) vempo uka koowi-ta me’a-k in like 1 and 2, there is no such
3PL the.SG pig-ACC.SG kill.SG-PRF requirement for the period.
“They killed the pig” The period is also only used
in line 2.
One of the great challenges of working on morphology are the ways that morphology
interact with other systems of grammar, particularly phonology, syntax, and semantics.
You might be coming to this book with considerable background in these areas, or you
might be relatively new to these areas. Regardless of your background, we will introduce
the necessary terms and concepts throughout this book in all related subareas of linguis-
tics – even if you are quite familiar with these other areas, pay attention to details we
build in this text; our study of morphology will cause us to revise and expand our
approaches to all of these other areas.
Let us examine some data from English pluralization to see how morphology inter-
acts with some of these other systems. To do this, we will use the International Phonetic
Alphabet (IPA) to represent the sounds of English. We will examine the IPA further in the
coming chapters.
Step 1 Observe: Our first job is to examine the data and to determine what the
morpheme(s) might be and what their distribution might be. Looking at the English data,
we should notice three different forms of the plural – each a suffix. The three forms are -s,
-z, and -əz. This will form the basis of our hypothesis generation.
Step 2 Hypothesize: As part of our hypothesis building, we will put some of the forms
we are examining into three-line, interlinear glosses. Recall that the claim that -s, -z, and,
-əz are independent morphemes is a hypothesis in itself. Our other goal is to try to deter-
mine anything we can about the contextual distribution to allow us to make predictions.
Step 3 Predict: As we examine our glosses, a clearer pattern should emerge. The suffix
-əz occurs when the final sound of the base it is attaching to is one of the set of sounds
{s, ʃ, t͡ʃ, d͡ʒ}. If we are previously familiar with phonology, we will note that each of these
sounds is a sibilant, which include the sounds {s, z, ʃ, ʒ, t͡ʃ, d͡ʒ}. Thus, we may make a test-
able about the distribution of -əz: “The suffix -əz occurs when the final sound of the word
it is attaching to is a sibilant.” Looking then at the distribution of -s, we should notice that
it occurs when the final sound has the following properties: it is both voiceless and non-
sibilant. Finally, looking at -z we should notice that it occurs when the final sound of the
base is both voiced and non-sibilant. We should then state these in testable statements.
Step 4 Test: For this step, we want to examine and gloss new but related examples. Our
goal is to see how our testable statements work in examples different from the ones we
originally examined. Because this is a sample exercise, the examples we choose will be
carefully constructed.
Thus, based on this data we could conclude that our statements were indeed valid
and stop. Of course, we have not tested all or even most of the English nouns.
Step 5 Observe: The key with good science is that we are never truly done, and we are
always looking for new examples to prove or disprove our account. In the case of English
plural, let us examine the following data:
In many cases, there are historical reasons for the irregular or suppletive status of a
particular morphological form, but we need to be very careful referencing such reasons
Chapter 1: Defining Morphology 11
in our explanations. Throughout this book, we are trying to build an explanation of how
the morphological component works in a first-language speaker’s linguistic system or
grammar. Our grammars are part of our mind-internal processes. What we learn about
our speech community’s language and our knowledge of morphology must come
through environmental exposure. Historical context is likely to be learned much later
than the irregular or suppletive forms; understanding the historical development of a
given word in a given speech community is information that is learned much later, if it is
learned at all.
This is not to say that studying historical context and historical change is not important.
The areas of historical linguistics and historical reconstructions are vital parts of the field
and have taught us much about the structure of language. Looking at the ways languages
change over time, analyzing the patterns, and reconstructing the earlier forms is known as
diachronic linguistics. Our primary focus is on the way speakers’ internal knowledge and
use of language presents in a given moment in time (typically in the p resent, though we
may also analyze older forms of a language) – such an approach is synchronic.
3. Theories of Morphology
It is our goal throughout this book to understand the rules and constraints that govern
how morphemes combine together. We will do this primarily through the framework of
Distributed Morphology (Halle and Marantz 1993). Linguists have developed a number
of different competing theories and frameworks for how the morphology part of the
grammar functions. Examples include Word-and-Paradigm Morphology (Blevins 2016)
and Lexical Phonology/Morphology (Kiparsky 1982). The focus of this book is on
Distributed Morphology, and we will not take the time to contrast this approach with all
of the alternatives. However, since most general introductions assume a simplified form
of Lexicalism or Lexical Morphology, we will contrast Distributed Morphology with
such an approach.
In a lexical approach to morphology, there is an independent component of the gram-
mar known as the Lexicon where words are stored and where morphological operations
occur. In many (perhaps even most) lexical theories of morphology, this lexicon is what
serves as the input to syntax. Within these frameworks, there is a need to capture the fact
that there is often overlap between what happens within the syntax and in the lexicon.
Which, if they were truly independent components, would not be expected. Principles
such as The mirror Principle (Baker 1985) restrict the types of operations that can occur
in the syntax and lexicon and force a correspondence.
If we examine how The mirror Principle functions, we will notice that it causes
a doubling of grammatical operations in a lexical approach. First, a morphological
12 Understanding Morphology
operation must occur in the lexicon. Then, an equivalent syntactic operation must
occur in the syntax. While this might ultimately be the correct solution, under the sci-
entific principles of simplicity of explanation, if there is a way to accomplish the same
results with fewer steps, then that solution should be preferred. That essentially is the
core of Distributed Morphology: to reduce complexity by eliminating the lexicon and
the associated redundancy of operations in it. In Distributed Morphology, syntax and
morphology are the same.
Distributed Morphology is generally situated within Chomsky’s Minimalist
Program. Key features of this program are the hypothesis that language is a “optimal
solution” and, as a consequence, a focus on economy. Many of these features fall
within the domain of the simplicity of explanation (modulo our assumptions). One
core development of the Minimalist Program is the so-called Y-model of syntactic
derivation.
Spellout
In the Y-model, the syntax pulls the items (words and morphemes) it will be manipulat-
ing and then applies morphosyntactic operations to them. At some point in this process,
the grammar decides to “spell-out” the sentence or phrase being produced. This sentence
or phrase is then sent to two different components of the grammar: P(honological) F(orm)
(where it is pronounced) and L(ogical) F(orm) (where meaning is interpreted).
Morphosyntactic operations can continue to happen on the paths to both of the final ver-
sions of PF and LF, so the pronounced form and the interpreted form often diverge from
each other.
You may or may not be familiar with this model of grammar – the details are not
immediately essential to us, so do not worry if this is Martian to you. We will cover this
model in much more detail in Part 2 of this book. But, it is important to consider the ques-
tions that this model raises for the framework of morphology. In particular, we should
ask ourselves what constitutes the “input” to such a model if we no longer posit a lexicon
where morphological operations occur. Such questions will remain with us for most of
this book.
The core ideas of Distributed Morphology may seem radical at first. They certainly
run counter to our basic intuitions about how language is organized and functions. As we
work our way through this book and develop the theory and framework together, we will
begin to see how these ideas obey the core scientific principles of simplicity of explana-
tion (Occam’s Razor) that were presented earlier in this chapter.
Chapter 1: Defining Morphology 13
4. Conclusions
In this chapter, we explored some challenges with defining the concept of the “word.” We
noted that there is no entirely consistent definition that perfectly captures the informal
notion of word that is used widely by linguists (and others). This presents an obvious
challenge when trying to understand the processes of word formation in natural lan-
guage. Despite this issue, we developed a first-look approach to morphology, which is
fundamentally a theory about word formation.
In this discussion, we examined how scientific investigation should proceed and how
we may utilize the elements of the scientific method to examine word formation in lan-
guage. We also introduced the core ideas of the framework, Distributed Morphology, that
we will be developing throughout this book. One critical aspect of Distributed Morphology
is that it distributes the processes of word formation across the syntax that allows us to
capture the significant overlaps found between syntactic and morphological processes.
Throughout this book, we will develop and refine this approach to word formation.
Further Reading
Aronoff, Mark and Kristen Fudeman. 2022. What is Morphology? 3rd Edition. Oxford:
Wiley-Blackwell.
Another introductory textbook on morphology that is designed to focus less on
theoretical approaches to morphology. Excellent resource for students looking for
additional support in the fundamentals of morphological analysis.
de Belder, Marijke and Jan Don. 2022. Distributed morphology: an oratio pro domo.
Nederlandse Taalkunde 27: 75–104.
Paper that explains the fundamental concepts of Distributed Morphology meant for a
nonspecialist audience.
Halle, Morris and Alec Marantz. 1993. Distributed Morphology and the Pieces of
Inflection. In The View from Building 20: Essays in Linguistics in Honor of Sylvain Bromberger,
ed. by Kenneth Hale and Samuel Jay Keyser. Cambridge, MA: MIT Press. 111–176.
One of foundational papers in Distributed Morphology. Advanced reading but a
good resource for understanding the reasoning and methodology in Distributed
Morphology. Consider rereading after completing this volume.
Harley, Heidi. 2006. English Words: A Linguistic Introduction. Oxford: Wiley-Blackwell.
Introductory and accessible text focused on English morphology. Develops a less tech-
nical version of Distributed Morphology to account for English morphology.
Haspelmath, Martin. 2011. The indeterminacy of word segmentation and the nature of
morphology and syntax. Folia Linguistica 45: 31–80.
A thorough breakdown as to why the concept of the “word” does not have clear
scientific definition.
Thorburn, William M. 1918. The myth of Occam’s Razor. Mind 27: 345–353.
A philosophy paper outlining some challenges to Occam’s Razor as a fundamental
concept in science.
16 Understanding Morphology
(A) koščïklar “little birds” (B) koščïklarnïŋ “of the little birds”
(C) koštan “from the bird” (D) košïbïz “our bird”
(E) balïklarïbïz “our fish (pl.)” (F) kaplar “bags”
(G) koščïk “little bird” (H) balïkïm “my fish”
(I) kapčïk “little bag” (J) balïkčïkïbïz “our little fish”
(K) balïknïŋ “of the fish” (L) kapčïktan “from the little bag”
(M) kapka “toward the bag” (N) balïklar “fish (pl.)”
(O) kapïbïznïŋ “of our bag” (P) balïkka “toward the fish (sg.)”
Part 1: Give the Tatar morphemes for the English equivalents of “bird,” “fish,” “bag,”
and “my.”
Part 2: Provide glosses for examples A, B, C, I, J, L, and M.
Part 3: Based on the evidence you have from Parts 1 and 2, provide glosses for these new
Tatar forms:
(Q) koščïkïbïz
(R) balïklarïmnïŋ
Part 4: Provided hypothesized Tatar translations of the following English phrases based
on the evidence you have gathered:
(S) “of my little bag”
(T) “toward the little bird”
Problem thanks to Karen Baertsch based on data from Poppe (1963).
Parse each form and provide a gloss for the forms found in the final two columns of the
table. Describe and explain any allomorphy found in the data.
Problem thanks to Sylvia Schreiner based on data from Jacobson (2001).
Part 2: Rewrite the above passage putting it in line with our discussion in this chapter.
Part 3: If available, use OpenAI’s ChatGPT (or a similar conversational model) to ask
the question “what is a word?”. Provide the response that you get and explain what the
response gets correct and what it gets incorrect. How does it differ from the response
provided here?
Part 4: Rewrite the new passage putting it in line with our discussion in this chapter.
Chapter 1: Defining Morphology 19
Now consider the data in Set II. This data exhibits yet a different pattern when it comes
to plural morphology. Describe it and discuss what it might mean for our approach to
morphology.
Set II
(A) “*pant (noun)” ~ “pants”
(B) “*short (noun)” ~ “shorts”
(C) “*scissor (noun)” ~ “scissors”
In plain terms, describe how the “intensive” meaning (the meaning associated with
continuing or successive action) is formed. You do not need to worry about the tones
(indicated with the diacritic marks on the vowels).
Data from Kraft and Kirk-Greene (1990 [1973]).