Another component of Language is morphology - literally, “shape”ology...

...The shape of what? Words/lexemes. Not the shape constituted by the sounds/phonemes - that would be phonology (“sound”ology), but the shape of the word as it is put together by minimal units of meaning (= morphemes, in contrast to phonemes).* A word itself may be just one morpheme - like the word “half.” A morpheme by itself is not necessarily a word, like the suffix “-ling”.# We can put the two morphemes together to make “halfling.” We could also take phoneme “a” - which by itself is also a morpheme! - and which by itself is also a lexeme/word! - and put it in front of “halfling”: “a halfling” (but that would be getting into syntax, for we would have transcended the word-level and entered the phrase-level).

    That’s morphology - concatenative morphology, to be exact. Not all morphology is about a linear string of morphemes put together like the cars pulled by a steam engine (which would be the governing head of the word). Sometimes, the morphology of a word can be more like a chameleon - whose body (the governing head) changes colors. An example of such a chameleon is the root “sng.” It has four colors: i, a, u and o. So, this chameleon has four manifestations: sing, sang, sung and song.

    Phonologically speaking, there are two types of morphology: there is concatinative “choo-choo train” hahaha morphology vs. non-concatinative “salamander” (^_~) morphology. The former tacks on phonemes (which themselves make up morphemes), the latter seems to alter the quality of the phonemes themselves. To illustrate the latter: One could think of the “sng” root as being really “sVng,” where V = vowel. But the nature of that vowel is itself the morpheme here; and the vowel’s nature is represented by features, not a whole phoneme; so the morpheme is a feature or group of features (but not so many features as to constitute a whole phoneme). Interestingly, there are whole languages based on consonantal roots, such as the Semitic ones. An example of a Semitic root is s(h)-l-m, from which we get the words “Salem, shalom, Islam, (mu)slem,” etc. This is not an isolated example - like I said, the Semitic languages are BASED on roots.

    English, in its ancient history even deeper back than Old English - all the way back to Proto-Indo-European (5,000BC?) - seems to have much of its vocabulary largely derived from roots. But, whereas the Semitic languages have TRIconsonantal roots, the Indo-European roots are BIconsonantal. For an example: gl. How many words do you know with gl that have to do with a shiny appearance? Gleam, glisten, glitter, glare, glow, glaze, glue, so on and on. Do all gl-words come from some Indo-European biconsonantal root? Or is this a case of some form of onomatopoeia? Dry coincidence? (To investigate this whole business about biconsonantal roots in English, just peruse an etymological dictionary, Bub.)

    A major characteristic of the Germanic languages is a type of verb, the so-called “strong verbs,” which have developed a powerful morphology out of certain purely phonological vowel-alternations in the biconsonantal roots of Indo-European. Aforementioned “sVng” is an excellent example of this. This vowel alternation - i.e, sing~sang~sung - is known as ablaut. (ng counts as one consonant here.)

    The history of ablaut begins in Proto-Indo-European with its biconsonantal roots. We shall take one root, wr, and trace it through time from Proto-Indo-European to Germanic, just to see the dynamics of the root and ablaut.......

Within Proto-Indo-European.......

root      wr


          wr + t → stem     wrt-

                                  ↙       ↘

                           wert-    ↔   wort-

                             ↕                   ↕

                          wērt-     ↔    wōrt-

NOTE: up and down axis from wrt without a vowel (zero mora) to having a vowel (one mora) to having a long vowel (two morae) is quantitative ablaut; the left and right axis from front vowel e to back vowel o is qualitative ablaut.

2ND NOTE: The ablaut patterns in Proto-Indo-European were purely phonological. The variation in the length of vowel (quantitative ablaut), and the variation in the timber of the vowel (qualitative ablaut), were two types of allophonic* reactions to shifting accent. Stress accent can effect the length of vowels, though Proto-Indo-European was supposed to have a floating musical (pitch) accent; pitch accent can effect the timber of vowels (pitch itself being “timber,” basically).

    In contrast, with the Germanic family of languages (an offshoot of Proto-Indo-European) the ablaut patterns became re-interpreted as morphological - which is to say, as carrying meaning...

    What we have in Proto-Indo-European as vowels e as opposed to o may have been the same vowel, let’s call it V, which alternated pitch-wise, e.g, V(pitch1) ~ V(pitch2). This timber gets re-interpretted as being a feature of placing (of tongue) in stress-based Germanic: e ~ o (tongue forward ~ back).

Within Germanic.......

NOTE: There are various characteristic sound shifts (Lautverschiebungen) when going from Proto-Indo-European to the Germanic family of languages (an offshoot of Proto-Indo-Euopean, as already stated).

We will use German here to exemplify Germanic...

Proto-Indo-European wVrt = wVrd in German...

werd(en), wird, so on (front vowels): present

ward (archaic), wurd(en), (ge)wurd(en) (back vowels): past

       With ablaut in Germanic we have taken a look at a productive morphological process involving alterations of features, really, and only non-directly of phonemes (vowels): thus, ablaut is a non-concatenative morphological process (i.e, when not purely phonological, as in Proto-Indo-European).........

    There is another great example of a single distinctive feature that all by itself is a morpheme to be found within the Germanic languages: the feature [+Front] being added to the root vowel 1) to make a nominus (i.e, noun or adjective) into a verb, or 2) to make a verb causative, or 3) with some nouns, to make them plural, or 4) often with verbs, to put them into the subjunctive mood. This sort of morpo|phono|logical (^_~) process is known as UMlaut (compare with ABlaut). Umlaut means to pronounce the vowel just as before but with the one exception of placing the tongue in the [+Front] position, which is to say in the position it would be in when one makes the “eee”-sound. Examples of 1-4: 1) German Zahl (number) > zählen (count) ; 2) sit > set, German sitzen > setzen; 3) man > men, German Mann > Männ(er); and 4) German war (= English was) > wär(e) (would be, were).

[NOTE: Just as with ablaut, umlaut (i.e, “fronting”) was originally a purely phonological process. The fronting of the root vowel was triggered by the appearance of a front vowel in the suffix: e.g, man + i (for the plural) => mani, which became meni which became just men. The suffix was lost, but the two forms, man and men, stayed.]

    Sometimes a morpheme may itself just be a prosodic pattern. Think of the following examples (capitalization indicating the stressed syllable): PROtest (noun) > proTEST (verb). Here we have an underlying initial stress (typical for Germanic), and then to make the noun into a verb, the prosodic pattern of unstressed-stressed is imposed on the lexeme. [NOTE: the verb PROtest (with initial stress) is a verb different from proTEST; the former means ‘to hold a protest,’ in the Sixties’ sense.]

    Just like there are two types of morphology phonologically (i.e, concatenative and non-concatenative), semantically there are two types of morphology, too. On the one hand, there is derivational morphology. It is the morphology involved in 1) making new words, and 2) changing the part of speech of words. The most straightforward example of (1) is compounding: super + man = superman. For (2): German noun Zahl > verb zählen. And we have already given other examples along the way. -On the other hand, there is inflectional morphology, which involves morphology that tends to be more “grammary” in nature - that is, on the one hand, a lexeme’s (grammatical) reaction to a particular syntactic instance (sentence) it finds itself in (e.g, case-endings in fusional languages such as Latin or Polish or Icelandic); or, on the other hand, a kind of tweaking of a less “grammary” or syntactic motivation, which would be to say something ranging closer to derivational morphology, something lexically based before the lexeme is inserted into syntactic instance (sentence)  - e.g, there is the plural s ending in English, and the ed past marker.

    Certain languages, like Latin, are notorious for having an absolute MESS (hahaha) of inflectional morphology - just look at a noun or verb paradigm in a Latin textbook. These paradigms are examples of inflectional morphology - in particular, that of a fusional language. But other languages, like Turkish, might have just as much inflectional morphology as Latin, with sizable paradigms as well, but there one could not say there is a “mess” of inflectional morphology: because in these languages the paradigms are not fusional buy agglutinative.

    Fusional* means every ending is not quite its own suffix, for through time its form has become fused with the stem of the word. This makes the paradigms messy and hard to learn, in the sense that there are no clear suffixes, just hints of certain suffixes that seem to take on almost as many forms as there are words that make use of them. Other languages, though, might keep their paradigms nice and neat with suffixes clearly being suffixes - this is with agglutinative languages, like Turkish.

    Morphology is very useful for classifying languages according to types. Agglutinative and fusional fall in a certain typological/morphological spectrum between isolative and polysynthetic. “Isolating” is the word for a language with no morphology. August Schleicher (see translation on this site) called this type of language “monosyllabic.” An ideal isolating language would have only lexemes and syntax with no morphology, and a simple phonology. On the other hand, an ideal/extreme (and unreal?) polysynthetic language would have no syntax but just morphology - no lexemes, just morphemes! “Polysynthetic” is the word for a language with such a huge amount of agglutination and/or fusion that, ideally, there would be no words, only affixes affixing to affixes HAHA! Or, it can be that there are only verbal lexemes, and everything else is an affix! The Native American languages are supposed to have this tendency. Notice how I said TENDENCY. (We are talking here about abstract ideals.) Schleicher did not take this type into account; Whorf talks about this type a lot in regards to his controversial theory of different languages determining how we perceive Reality - and arguing that Native American languages had more to do with real Reality than did the Indo-European languages. Schleicher considered, first, monosyllabic, then agglutinative, and then fusional, saying that one evolves into the other (monosyllabic to agglutinative), which then evolves into the next (agglutinative to fusional). The next evolutionary step would be polysynthetic. But, like I said, he did not go that far. To me, it seems an ideal isolating/monosyllabic language would have just nouns - they would become verbs temporarily through the appropriate context in the sentence; whereas polysynthetic languages, ideally, would have just verbal morphemes. Nominal lexemes on the one hand (isolating); verbal morphemes on the other (polysynthetic)... And what would come after polysynthetic? A language where the morphology and phonology are one and the same? Where every distinctive feature is a morpheme? Is that a root-based language like the Semitic? - a different type not discussed as a set type by Schleicher? Hmmm... And then, for the sake of completeness, let us ask what happens if we took this type of language where Morphology=Phonology (the morphophonological type) to the extreme/Ideal... Would not that be the ultimate language, some sort of Superlanguage to make Leibniz and Russell simply drool - and yet, at the same time, would not that be a reversion to animal cries...? (And would there be a difference?)

{{{Interesting to note here, in the root-based language Hebrew, with Hebrew  mysticism, i.e, the Kabbalah, each phoneme is ascribed a meaning, in the sense of a number...}}}


   August Schleicher [see here] said there are three classes of languages: one he calls "monosyllabic," the other agglutinating, the other fusional. There is nothing surprising about this, however; this is a well-known way of classifying languages, though nowadays we would add "polysynthetic" to that list, and instead of "monosyllabic" would use the term "isolating" . And let’s also add root-based.

monosyllabic/isolating - example, Chinese

agglutinating - example, Turkish

fusional - example, Latin

polysynthetic - example, Aztec

root-based - example, Hebrew

    The first three, which are the three Schleicher has, are based on the ideas of lexical meaning vs. grammatical relation. Lexical meaning is, basically, the definition of a word. Let's take the word "dog." It has a definition, and we know what it is. (Look "dog" up in a dictionary if you don't know what it means, hahaha.) Now, we can add "s" to the end of "dog" - like so: "dogs". That makes the word plural. The "s" in dog means PLURAL, or, to put it another way, it designates the grammatical category of plurality. So, "dog" has a lexical meaning, and the "s" ending bears the grammatical relation...

    In an isolating/"monosyllabic" language, a word would not have a suffix or ANYTHING to denote anything grammatical: "dog" could be singular or plural or WHATEVER in an isolating language. Likewise, in an isolating language, there would be nothing equivalent to our “ed" ending (which we have in English for most verbs to denote the grammatical category of past tense). Whereas in English we have, for example, "kiss" (present) as opposed to "kiss" + "ed" = "kissed" (past) - in an isolating language there would be just "kiss" for present or past or WHATEVER!!! ;)

    Stringing forms together - as in, kiss + ed = kissed - in a way where the boundary between each element is perfectly clear - kiss/ed - is what an agglutinating language does. "Agglutinating" literally means "gluing together."

    In an isolating language, the lexical meaning is expressed in sound, but the grammatical category is left up to context: dog can be singular or plural, subject or object; kiss can be past or present, singular or plural. In an agglutinating language, the lexical meaning is expressed with one element - that is, the word - and then the grammatical categories are expressed with other elements, that is, prefixes or suffixes (basically): so we have dog for the singular but dog + s = the plural; and kiss for the present but kiss + ed for the past. s and ed were AGGLUTINATED (glued onto) dog or kiss, respectively.

    In a fusional language, the lexical meaning and the grammatical category/relation are expressed AT THE SAME TIME. Theoretically, this would mean that for every possible grammatical context of a word, it must have that form in its lexicon (=what a layperson would call "the dictionary," though we are not talking about the physical dictionary or any book here, but a real aspect of our unconscious minds). Here is an example of this from English: go = present, and then the totally different word "went" is used for the past. Now, technically that is not fusion but suppletion (the replacement of one form for another); however, this suppletion demonstrates the wholeness of form that a fusional language is supposed to adhere to.

{{{This shows how fusional languages can develop into isolative - i.e, the inflected forms end out lined up in the lexicon... Eventually, the lexicon is overfull of words with no semantic differences but just grammatical - hence, the lexicon purges itself in the form of what is known as “leveling,” i.e, when a language gets rid of its case-system, like Latin became thus ever more heavy-laden with its grammar until it just dropped it with a sigh of relief giving way to the Romance languages...The lexicon thus mopped itself up.

The tree of Latin had leaves that grew and grew,

In collections upon collections of colors and shapes,

And became so heavy that this tree began to droop -

Until the autumnal wind blew them away -

A brief moment of naked winter,

Then Romantic Spring did arrive! }}}

    You may have seen already that I have given examples from one language for each of the three examples. It seems to be the case that every language has all these linguistic types present in them to a greater or lesser extent; and which linguistic type predominates in a given language, is that type of language that it is classified to be. {{{Assuming doing so would be useful... I mean, how much are we not perhaps talking about those illusory semantic circles around entities that are just as much connect to everything else in the universe as to themselves - like the aspects of the Self according to Zen in the Prajna Paramita Heart Sutra, and the erroneous word “race”???}}}

    The last type - one that Schleicher did not talk about as its own distinct type - was that of polysynthesis. Another word for this type of language is incorporating. But the two terms do not mean the same thing. Polysynthetic basically means that a whole slew of things - a veritable $H!TLOAD :) LOL - can be agglutinated onto the base word. There may or may not additionally be fusion, such as certain suffixes, for instance, being fused with many, many grammatical categories - such as by having one suffix that indicates both the subject and object of the verb: this is to be found in languages of the Americas. The base word is typically a verb. So many elements may be agglutinated to the base word, that often that one word - which is most likely a verb - would equal one entire sentence. And other words - not just grammatical affixes - may be compounded to that verbal base - such that the one core word (verb) may be equivalent to not just a sentence, but one large complex sentence. This process whereby many words are compounded to a certain base word is incorporation - which literally means a word bringing other words into its body (inCORPorate comes from Latin CORPis, which means "body"). We could construct a verbal example of incorporation in English: "meat-eat" (eat meat), "dish-clean" (clean the dishes). [See §17: ‘Framing the sentence’, pp. 128-139 of On Language: The Diversity of Human Language-Structure and its Influence on the Mental Development of Mankind by Wilhem von Humboldt (translated by Peter Heath), 1989, Cambridge University Press, Great Britain]

    When Schleicher made his threefold class distinction, he based them on a spectrum where, at one end, the given language did not involve the base word in any reference to grammar; and then, the intermediate stage, where a grammatical element was attached to the word; and then, in the final stage, where that grammatical element was merged with the word. Another way to view this spectrum: 1) there is word-unity but no grammatical expression; 2) there is no word-unity but there is grammatical expression; 3) there is both word-unity and grammatical expression. Polysynthesis/incorporating would exist as 2 or both 2 and 3 - it does not fit into Schleicher's spectrum very well at all, and, indeed may even make one question the validity of the entire spectrum - don’tcha think?

    Schleicher, coming from a long time ago, regards 3 as some sort of creamy evolutionary ultimate, as though all languages were in some sort of mystical evolution towards some sort of “superior” mode of talking. One must keep in mind that he comes from Old School Western culture, which had a fetish for "Classical" culture, i.e, that of the Romans and Ancient Greeks, who indeed did speak with 3 (fusional languages).

Schleicher’s Parameter’s of Language Type

language              preservation     perservation

type                     of word-unit     of relation

                           (Worteinheit)     (Beziehung)

“monosyllabic”*        yes                    no

compared to crystal

agglutinating             no                     yes

compared to plants

fusional                      yes                    yes

compared to animals

  1. *We could also use the term “isolating” .

Compare the above table with the following - a more or less table, or list, which, basically, is my interpretation of the prose of Bernard Comrie. pp. 42-52 (“Morphological Typology”) of Language Universals and Linguistic Typology (1989, University of Chicago Press)...

Parameter I... in the form of a spectrum...

having syntax,

but no morphology*          =     isolative


having syntax                     = agglutinative

and morphology                     or fusional


no syntax                           = polysynthetic

but having morphology#, %

* has only lexemes, no morphemes

# has only morphemes, no lexemes

% By definition, this would make the language an incorporating language - a term often used synonymously with “polysynthetic” - though Comrie wisely regards incorporation as separate from polysynthesis: “Although the two terms are sometimes used interchangeably, it is possible and advisable to make a distinction between them.” [pg. 45] This is because the two terms do represent two distinct morphological processes: polysynthesis simply means an enormous amount of morphemes can be tacked on to the root or base of the word or lexical morpheme; incorporation means words or lexical morphemes are all put together into one big word which, in its extreme, may equal an entire sentence. However, in our extreme/ideal definition of polysynthesis with no syntax but with morphology, every sentence would really be one word (there would be no syntax to handle more than one word in a sentence): this word would be an incorporation of morphemes. But this is only an ideal. It is true that incorporation can take place in areas of the spectrum other than polysynthetic langugaes, e.g, compounds are a type of “light” incorporating.

Parameter II (two dimensional)...

    1)  monosyllabic       ~     polysyllabic

    2)  concatenative     ~  non-concatenative

          morphology               morphology

Imagine an X over the four possibilities: the X has two lines and the line going from “monosyllabic” down to “non-concatenative morphology” would be exemplified by Hebrew; and the line going down from “polysyllabic” to “concatenative morphology” would be exemplified by Turkish.

Parameter III...

This is a parameter just for polysyllabic languages with a morphology*. It is the question of whether or not the boundaries of morphemes and bases (in the sense of Schleicher’s Beziehungslauten and Bedeutungslauten) are clear...


morphemic               = agglutinating

boundaries                   (e.g, Turkish)

unclear                     = fusional

morphemic                  (e.g, Latin)


    We can think of morphology from the ground up, just like we did with phonology, beginning with the smallest, most basic and fundamental unit, and work our way up the rungs of the morphological ladder to increasingly larger, more category-encompassing and general units. We can start by taking the following biconsonantal Indo-European root as it appears in German: lb or lVb (with V standing for any vowel), depending on how you may wish to think of it...

1) root lVb becomes the following stems...

     (this is derivational morphology)

     l[ie]b               or          l[au]b

     (this is non-concatinative morphology)

2) stem lieb- becomes the following lexemes...

     (this is derivational morphology)

     lieb + e           or          lieb + en

       noun                              verb

      “love”                          “to love”

  1. 3)stem laub- becomes the following lexemes...

     (this is derivational morphology)

      g(e) + laub + en       or      er + laub + en

            v  e  r  b                           v  e  r  b

         “ to believe”                     “to allow”

  1. 4)inflectional morphology...

    liebe (noun)

        remains liebe (spelt Liebe) when singular,

        becomes liebe + en (Lieben) when plural

    -en infinitive ending dropped & replaced

    by the following in the finite forms...

     [tense ending] + [person/number ending]

    ∅ for non-past       -e      1st person sg.

     -t for past             -(e)st   2nd person sg.

                                   -t      3rd person sg.

                                  -en  1st & 3rd person pl.,

                                          & honorific 2nd

                                  -et     2nd person pl.

  1. 5)for the participle forms,

       infinitive suffix -en replaced with...

         -end for the

            present/active/progressive participle

         -t for the

            past/passive/perfect pariticple

        NOTE: -t actually part of circumflex ge---t (fits around stem), but there is no place for ge- to appear if stem has an inseperable prefix (e.g, the g(e)- & er- in g(e)lauben and erlauben).

    Indeed, there is the same two-sidedness to the morphological ladder that there is to the phonological ladder: on the one hand there are the minimal units in regards to linear, unit-by-unit strings of elements; on the other hand there are the minimal units in the sense of prosodic templates* or abstract features# applied to the linear strings. For this reason, in phonology one can just as much consider a certain prosodic unit - like mora or syllable or foot, as we’ve discussed in the phonology section - as the most minimal unit recognized by the grammar, as opposed to linear units like phonemes. Because morphology must make use of sound, the two-sidedness of the sound component of grammar, i.e, phonology, applies equally to morphology.


The evolutionary path of grammaticalization is on the top, the corresponding evolutionary path of linguistic types is on the bottom...

lexeme > clitic >  affix > fused element

isolating > agglutinating> fusional


  1. *PHASE I:        lexeme + lexeme


Examples (from the “dental” or “tongue tip” preterit [=simple past tense] in Proto-Germanic):

habe dedum     (literally: ‘have did [we]’)

habe deduth     (literally: ‘have did [ye]’)

habe dedun      (literally: ‘have did [they]’)

  1. *PHASE II:        lexeme + clitic

  (less isolating, somewhat agglutinating)

Examples (second element pronounced less distinctly):




  1. *PHASE III:        lexeme + affix






  1. *PHASE IV: one “flexing” lexeme!!! (^_^)


Examples: hadd- (past) ~ hab- (present)


cognates:   had                  have

(NOTE: b in hab pronounced much like v)



Contents of this lecture:

  1. 1)Introduction

  1. 2)Phonology

3) Lexemes

4) Morphology (you are here)

  1. 5)Syntax

  1. 6)Conclusion

# A morpheme that is free to stand alone as a lexeme is a free morpheme - as opposed to a morpheme that can only be used when connected to a lexeme or other morpheme; the latter is a bound morpheme.

* Phonemes - or sometimes, as we have seen in the phonology section - just plain features in themselves - are the minimal units of sound (sound as opposed to meaning, which is to say form as opposed to substance): but not the minimal unit of sound physically/objectively speaking, but in terms of what the grammar of a given language recognizes (see Introduction).

* For example: imposing the stress pattern of unstressed-stressed onto nouns to make them verbs.

# For example: umlaut, i.e, the feature of [+Front] added to a vowel.

* In a sense, what “allophonic” means, in part, is that it was purely phonological,  without semantic meaning, and involving sound changes which, according to the grammar, were not “supposed” to be heard/heeded.

* NOTE: We have not considered polysyllabic without a morphology - but why could there not be this type? Schleicher seemed to say there was a natural reason, which was simply “the unity [Einheit] of the concept, the idea is reflected in a unit [Einheit] of sound (a syllable).

* See an example of a language with both fusional and agglutination in its morphology with my artificially constructed langugae Atlantian - click here.

For another illustration of ablaut and root morphology in Indo-European and Germanic, please click here.Roots.html