The process of agglutination may be accomplished in many different ways, any of which may be characteristic of whole groups of unrelated languages. These may be roughly divided first into mono- or oligo-synthetic and polysynthetic. The former very nearly approach the isolating languages, since usually only one element may be added at a time, but the process of addition may be accomplished in any of the ways possible to agglutination.
Agglutination includes prefixing, suffixing and infixing in all degrees of complexity and fixity. Thus languages may be spoken of as agglutinative only in a relative sense. Some are much more so than others, both in point of the number of elements which it is possible to add, and their dependence upon one another and the root, denoting a higher or lower degree of inextricability in blending.
Many languages are only loosely agglutinative and the component parts of the compounds readily resolve. In others, as in the inflecting languages, the combination is inextricable.
Thus under the head of agglutinative we have the merely agglutinative or synthetic, readily resolvable combinations, which are often hardly distinguishable from isolating languages, and the less easily divisible inflectional and incorporating types. Any or all of the three processes of infixing, prefixing and suffixing may be employed in simple agglutinative combinations.
In inflectional languages the root is attended by prefixes or suffixes which form inseparable modifiers. At times phonetic changes occur which render the complex unlike the simple joining of its component parts.
As Mr. Sweet says: “If we define inflection as ‘agglutination run mad’ we may regard incorporation as inflection run madder still, for it is the result of attempting to develop a verb into a complete sentence.” In some languages, such as the incorporating, a verb is sufficiently distinct in its meaning not to require an independent pronoun. French and Spanish, though not belonging to this category, contain words with the incorporating idea, as in Spanish hablo, I speak, and French, pluit, it rains. Where polysynthesism is the prevailing character, the verb may be sufficiently comprehensive to include the objective pronoun as well as the subjective, so that it is possible to find in one word a transitive, as well as in others an intransitive, sentence. But this is only rudimentary incorporation, and borders on inflection. Some American Indian languages carry it to a very high degree, appending to or inserting into this simple complex not only nouns which may stand in apposition to the implied or actual pronouns, but particles and modifiers of every description. (See the Handbook of American Indian Languages, published by the Bureau of American Ethnology at Washington.) Frequently during this process various parts undergo phonetic changes in accordance with fixed laws, so that the final complex may not at all resemble a string of the original elements, but becomes a new, inseparable and fixed word containing a whole sentence of ideas. This sentence, in some languages, may carry throughout certain modifiers for all noun elements—for instance, as to whether the objects under discussion are visible or invisible. These modifiers bear definite relationships to the nouns, and the “sentence word” in each of its parts must then be conjugated as a verb in an even more complicated manner. This is agglutination par excellence, and is frequently so complex as to be utterly bewildering to the Indo-European mind, even though the Indo-European languages themselves employ agglutination to a limited degree and of certain varieties, particularly of the inflectional order.
Compared to the most complicated Indian tongues, English is in the position of Chinese to Indo-European languages in its structural simplicity, though of course in Chinese we have an added complexity in the use of pitch, etc.
There are certain types of speech which secure changes (plurals, etc.) by internal vowel modification. English itself makes use of this device, but it is the outstanding feature of Semitic tongues.
Sweet says: “There are many other minor criteria of morphological classification. The most important of these is perhaps that of the agglutinative or inflectional elements before or after the word or stem [modified]. In Turkish and in other Altaic languages, as also in Finnish, these are always post-positions, so that every word begins with the root which always has chief stress. The Bantu languages of South Africa, on the other hand, favor prefixes.... The Semitic languages favor prefixes and post-positions about equally. The Aryan languages are mainly post-positional, with occasional use of prefixes, most of which, however, are of later origin.”
It must not be supposed that languages fall into absolutely distinct categories because of their structure. No language to-day is purely of one type or another. There have been too many centuries of borrowing and change for that condition to now be possible for any language, nor are there any longer what might be called primitive tongues. They have all long since outgrown that state, whatever it may have been, even the Botocudo of Brazil, which is generally ranked as the most primitive.