3.3.11 Affixation and compounding
In general, variation in lexical morphology has been recorded in the tagging somewhat less systematically than grammar. For example, nouns ending in -ion (e.g., information, obligation, observation, etc.) have not been analysed, whatever their history in English. This is because a number of alternative data-retrieval methods are available for studies of word-initial and -final elements.
In fact, elaboration on features of lexical morphology focuses on derivation by affixes and compounding. As discussed in Section 3.2, the user should take into account the following decisions, which may negatively affect the coherence and consistency of the present tagging system. The morpheme -ly in open-class adverbs has been tagged as a suffix (xs-av). Similarly, comparative and superlative morphemes are tagged as /xs:
$strong/aj-cpv_STRONG+ER $-er/xs-aj-cpv_+ER
$fit/aj-sup_FET+TEST $-est/xs-sup_+TEST
The element -self in reflexive pronouns has the grammel /xs-P:
$/P23X_YAm-SELF+Fis $-self/xs-P_-SELF+Fis $/P23X_+Fis
The main rule is that lexical morphemes are tagged separately when there is evidence for them being productive in English; the Dictionary of the Older Scottish Tongue (http://www.arts.ed.ac.uk/dost/) and the Oxford English Dictionary (http://dictionary.oed.com/) were regularly consulted in the decision-making process. However, there are a couple of exceptions to this rule. For example, derived adjectives with the suffix -(i)ous and derived nouns in -ment have been analysed separately, even though these adjectives and nouns are usually direc t borrowings from French and as such should be left unanalysed in the tags. The practice has been adopted because of the tagger's special interest in the variation attested in these elements; the otherwise fairly stable -ment shows variation between contracted and full forms, the former indicated by a flourish.
$desire/aj_DESYR+US $-ous/xs-aj_+US
$gracious/aj_GRAC+IOUS $-ous/xs-aj_+IOUS
$affectuous/av_EFFECT+UOS+LY $-ous/xs-aj-av_+UOS+LY $-ly/xs-av_+LY
$marvellous/av_MerWEL+WSs $-ous/xs-aj_+WSs $-ly/xs-av_0
$advancement/n{rc}_AUAnC+EMAnT $-ment/xs-n{rc}_+EMAnT
Similarly, nouns with the final morphemes -er, -or and -ar and their variants have been tagged as suffixes even though only the first morpheme occurs frequently in deverbal derived nouns which are transparently English coinages. Thus, the instances of affixation in words such as bearer, governor and servitor are positioned in the same variational pattern through the use of tagging, even though the latter two are loan formations:
$bear/n_BEAR+ER $-er/xs-n_+ER
$governor/n_GOUERN+Our $-or/xs-n _+Our
$servitor/n_SerUIT+ur $-or/xs-n_+ur
Prefixes are tagged with the core property element /xp, negative ones with /neg-xp:
$carry/vpp{pass}_MIS+KARI+ED
$mis-/xp-vpp{pass}_MIS+ $/vpp{pass}_+ED
$kind/n{rc}-av_VN+KYND+NES $un-/neg-xp-aj-n{rc}-av_VN+ $-ness/xs-n{rc}-av_+NES
$/S_jT
{\}
$be{n}/vps13<S+_jS
$/neg<v>neg-xp_NO^T
$know/venpp{pass}-aj>pr_WN+KNAWINE $un-/neg-xp-venpp{pass}-aj>pr_WN+_
$to/pr+C<venpp-aj_TO
$/P02G_zO^R
$lordship/n{ho}_L
$say/vpp{pass}-aj{post}_FOR+SAID $fore-/xp-vpp{pass}-aj{post}_FOR+
In native formations such as the verb type overthrow, the first element is analysed as an adverb in a verb compound, as this practice permits the study of alternation between overthrow and the phrasal verb throw over:
$over/av>vpt-k_OUR
{\}
$throw/vpt-k<av_THREW
As the above example shows, the analysis of over as a prefix is not supported by the spelling as two words. Similarly, the first unit in the noun indweller is also tagged as an adverb:
$in/av>npl-k_IN+
$dwell/npl-k<av_+DUELL+AR+IS $-er/xs-npl-k<av_+AR+IS $/pln-k<av_+IS
The core property symbol for suffixes is /xs:
$bear/n_BEAR+ER $-er/xs-n_+ER
$honour/aj_HONOR+ABiLE $-able/xs-aj_+ABiLE
$forget/aj-n{rc}_FORGETT+FULL+NESS
$-ful/xs-aj-n{rc}_+FULL+NESS $-ness/xs-aj-n{rc}_+NESS
$trouble/aj_TROUBLE+SUM $-some/xs-aj_+SUM
$will/vpsp-aj-n{rc}_WILL+ING+E\NES $/vpsp-aj-n{rc}_+ING+
$-ness/xs-vpsp-aj-n{rc}_+E\NES
$goodly/aj_GWD+LY $-ly/xs-av_+LY
There may be comments highlighting a particular function, such as the use of a derivative as an honorific:
$/P13GM_HIS
$high/n{ho}_HIE+NES $-ness/xs-n{ho}_+NES
$/P02G_zOur
$ladyship/n{ho}_LADY+SHIP $-ship/xs-n{ho}_+SHIP
Units in compounds have also been tagged separately; the head of a compound is marked with the property -k and arrows indicating concatenation:
$/T_THE
$up/aj-cpv>n-k_WP+PER $/aj-cpv>n-k_+PER
{\}
$house/n{coll}-k<aj-cpv>pr_HOUSE
$of/pr<n-k_OF
$parliament/n_*PARLIAMENT
$/P02G_zOur
$own/aj_AWIN
{\}
$here/av>vn-k_HEIR+
$come/vn{rc}-k<av_+CUm+ING $/vn{rc}-k<av_+ING
$by/av>vpp-aj-k_BY+
$go/vpp-aj{post}-k<av_+GAINE
$well/av>vpp-aj-k_WEILL
$content/vpp-aj-k<av_COnTENT+IT $/vpp-aj-k<av_+IT
$weapon/n>vnpl-k_WAPPYn+
$show/vnpl{rc}-k<n_+SCHAW+YnG+is $/vnpl{rc}-k<n_+YnG+is $/plvn{rc}-k<n_+is
$evil/aj-n>npl-k_EUILL+
$will/npl-k<aj-n_+UILL+AR+S $-er/xs-npl-k<aj-n_+AR+S $/pln-k<aj-n_+S
$annual/aj>npl-k_Annual+
$rent/npl-k<aj_+RENT+S $/pln-k<aj_+S
Compounds in premodification are indicated as follows:
$north/aj>n-aj-k_NORTH+
$land/n-aj-k>npl-k<aj_+LAND
$man/npl-k<n-aj-k_MEN~
$new/aj>n-aj-k_NEW
$year/n-aj-k<aj>n-k_zEIR
$mass/n-k<n-aj-k_MESs
The pattern is intended to reflect the tagging of French men:
$french/aj>npl-k_FRANCH
{\}
$man/npl-k<aj_MEN
Compounds functioning as names are marked with the comment {onom} 'onomastic (use)' attached to the grammel:
;_*DANMARKE $/n>n-k_*DANMARKE
$house/n{onom}-k<n_HOUS
$/T_YE
$low/aj{onom}>npl-k_*LAW+
$country/npl{onom}-k_+CONTRI+ES $/pln{onom}-k_+ES
Instead of being analysed as a compound, phrases such as son-in-law have been tagged as a collocate:
$son/n>pr_SOUNNE
$in/pr<n>n_IN
$law/n<pr_LAW
Degree of lexicalization has not been considered in the tagging practice, but even in clearly lexicalized compounds the elements have been tagged separately:
$good/aj>n-k_GOUD+
$father/n-k<aj_+FATHER
$gentle/aj>npl-k_GENTIL+
$man/npl-k<aj_+MEN~
$well/av>n-k_VIL+
$fare/n{rc}-k-av_+FAIR
$grand/aj>n-k_GRAND+
$child/n-k<aj_+CHILD
$wild/aj>n-k_WILD+
$fowl/n{coll}-k<aj_+FOULLE
It should be noted that any nominal in premodification is interpreted as the first element in a compound, and it is possible to infer the compound type by analysing the context (e.g. 'at the exchequer and the council' suggests a locative and 'related to, concerning the exchequer and the council' an associative reading):
$/P11G+V_MY
$exchequer/n>n-k_EXCHECKER
$&/cj_AND
$council/n>n-k_COUNSELL
$business/n-k<n&n_BUSSI\NES
|