Morphosyntax

As we have seen when we discussed morphology on the previous page, and especially in conjunction with parts of speech, often words only tend to reveal their meanings in context. Additionally, we can often change existing words, so that they either take on a new role or reflect the ‘interaction’ of the individual word classes in a ‘sentence’ context. The former is mainly achieved by the word-formation processes of derivation and compounding (and to a somewhat minor extent also through back-formation, clipping & blending, and acronym formation), and the latter through inflection. We will discuss the roles of both of these major process classes in more detail below.

The Role of Derivation & Compounding

Derivation and compounding are primarily used in order to create new word forms involving a change in lexical meaning, as well as often also to change the function of the original word in a given context. As we have already discussed, deriving new word forms by attaching prefixes or creating new compounds usually has a kind of ‘diversifying’ effect, mainly preserving the original word class, whereas suffix derivation tends to cause a change in word class.

Thus, from the perspective of (morpho-)syntactic function within the larger linguistic unit, it does not matter whether we say This issue is important vs. This issue is unimportant or I ate all the cream vs. I ate all the ice cream; it is only on the level of meaning that we perceive this difference, as both important and unimportant equally function as adjectives, and cream and ice cream as nouns. The respective ‘prefixes’ do not cause any change in the functions the word forms fulfil; neither do they influence the overall structure of the linguistic unit. We can perceive a slight difference, though, if we use certain prefixes with verb stems in order to diversify their meaning. Therefore, while e.g. both receive and conceive have the same stem/root {ceive} and both do not change their function as verbs because of the prefigation process, receive more or less evokes only one contextual pattern, i.e. that of an object which is received, whereas conceive shows a greater range of potential contexts/constructions it may evoke (examples taken from the British National Corpus; BNC):

On the other hand, if we e.g. derive the word beginning from begin by adding the suffix {-ing}, we change the function of the derived word form from a verb to a noun. The following two sentences – taken from the Freiburg-LOB Corpus of British English (FLOB) – illustrate the difference:

The Role of Inflection

Inflection, on the other hand, never changes the word class of a word form, but instead always signals a change in its grammatical function (along with its meaning in context).

Most inflectional functions in English are achieved via suffixes:

In relatively rare cases, some inflectional processes may be reflected through a change in the stem vowel – or more abstractly – the use of the same form for singular & plural. This happens especially with strong/irregular verbs and nouns. Examples of this are singular-plural pairs of nouns, such as goose-geese, foot-feet, sheep-sheep and irregular present—simple past—past participle paradigms for verbs, such as put-put-put, give-gave-given, run-ran-run, etc.

In many inflecting languages, such as Latin or German, the grammatical function of nouns is (at least to some extent) signalled by case marking suffixes. Older forms of English – due to its Germanic origin – also had this feature, but in modern day English (which has become more analytic), case marking is restricted to such word classes as pronouns and question particles. As a ‘by-product’ of case marking in pronouns, we also tend to find gender marking in the third person singular forms.


Exercise

Look at the list of case-marked items above and try to identify in which way English differs from other languages, such as German, Latin, Russian, Greek, etc., that employ case to indicate grammatical functions. In which way may there be a gap in English, and how can it be filled?


‘Constituent’ Structure & Internal Ordering

Although we have already discussed the role of affixes in the word-formation process and with regard to their functionality in larger units, we have not yet discussed the option of multiple morphological processes occurring in a row in order to derive complex polymorphemic words. Obviously, by definition, prefixes get attached in front of a stem and suffixes at the end, but what happens when we want to attach more than one pre- or suffix? Well, this is where the role of constituent ordering of the individual morphemes comes into play. In order to illustrate that some morphemes may be attachable to others in different positions while others are not, we’ll do two further exercises.


Exercises

  1. Split the following word forms into maximally three parts and paste them in the three text boxes on the form: unconventionality, disloyal, manageableness, unhealthy, pretreatment. Click the scramble morphemes button repeatedly and determine whether the resulting word form does make sense in any way. Please note that, as the ordering is randomly produced, you may sometimes get the same form, also including the original word, twice.
  2. Using the following list of pre- and suffixes (+ possibly your own examples), try to find stems they may combine with and test potential combinations/permutations in the form below (without doing the scrambling): {post-}, {pre-}, {re-}, {counter-}, {cross-}, {forth-}; {-ity}, {-ism}, {-ness}, {-tion}, {-sion}, {-er}, {-ess}; {-able}, {-ly}, {-ish}, {-ful}, {-ous}; {-nate}, {-iate}, {-en}, {-ing}, {-wise}. Try to identify the functionality of each of the affixes and reflect upon why they may only occur in certain positions.
{}{}{}

As we will see later, similar conditions hold for the ordering of constituents on the level of syntax.


Morphosyntactic Annotation

Morphosyntactic annotation or tagging is a process whereby a part-of-speech label (tag) is associated/attached with each word (or punctuation mark) in a text in order to reflect its function. This process is usually carried out more or less automatically using a computer program called a tagger, although, of course, purely manual annotation is also possible, albeit very time-consuming. You will normally encounter (at least) two different forms, a slightly older one where the tagged word is followed by an underscore character (_) and a grammatical tag, such as e.g. house_NN1, or the more modern version where the word itself is enclosed in some kind of SGML or XML element, such as <word pos="NN1">house</word> and the part-of-speech tag appearing as the value of the pos attribute. Examples of well-known tagsets used for morphosyntactic tagging are CLAWS (Constituent Likelihood Automatic Word-tagging System; version C7; version C5) and the Penn tagset.


Exercise

The following two sentences are the tagged versions of the two examples from the FLOB corpus presented above. In order to develop your analysis skills, try to find a suitable strategy to help you recognise regular patterns that allow you to identify what each PoS tag stands for, paying particular attention to the word-formation processes that may give rise to the tags. At the same time, try to develop some ideas as to why this kind of tagged text may be useful for grammatical analysis, and what the distinctions in the tagset may tell us about the need for more or less fine-grained descriptive categories. Use the box below to write down your observations and develop your strategy. Hint: Try grouping related categories together.

People_NN begin_VV0 to_TO spend_VVI again_RT ,_YCOM which_DDQ means_VVZ the_AT economy_NN1 begins_VVZ to_TO grow_VVI and_CC there_EX is_VBZ a_AT1 general_JJ virtuous_JJ circle_NN1 ._YSTP
Most_DAT governments_NN2 lose_VV0 support_NN1 during_II election_NN1 campaigns_NN2 ,_YCOM although_CS Mrs_NNB Thatcher_NP1 broke_VVD the_AT pattern_NN1 in_II 1987_MC by_II remaining_VVG steady_JJ from_II beginning_NN1 to_II end_NN1 ._YSTP

Once you’ve tried to come up with your own solution, you can to see what your initial strategy could have looked like. And if you want to explore tagging itself a little further, you can try to use the CLAWS trial service online to experiment with tagging your own short texts.