CHAPTER XXXI.

COMPOSITION.

s. 410. In the following words, amongst many others, we have palpable and indubitable specimens of composition. _Day-star_, _vine-yard_, _sun-beam_, _apple-tree_, _ship-load_, _silver-smith_, &c. The words _palpable_ and _indubitable_ have been used, because, in many cases, as will be seen hereafter, it is difficult to determine whether a word be a true compound or not.

Now, in each of the compounds quoted above, it may be seen that it is the second word which is qualified, or defined, by the first, and that it is not the first which is qualified or defined, by the second. Of _yards_, _beams_, _trees_, _loads_, _smiths_, there may be many sorts, and, in order to determine what _particular_ sort of _yard_, _beam_, _tree_, _load_, or _smith_, may be meant, the words _vine_, _sun_, _apple_, _ship_, and _silver_, are prefixed. In compound words it is the _first_ term that defines or particularises the second.

s. 411. That the idea given by the word _apple-tree_ is not referable to the words _apple_ and _tree_, irrespective of the order in which they occur, may be seen by reversing the position of them. The word _tree-apple_, although not existing in the language, is as correct a word as _thorn-apple_. In _tree-apple_, the particular sort of _apple_ meant is denoted by the word _tree_, and if there were in our gardens various sorts of plants called _apples_, of which some grew along the ground and others upon trees, such a word as _tree-apple_ would be required in order to be opposed to _earth-apple_, or _ground-apple_, or some word of the kind.

In the compound words _tree-apple_ and _apple-tree_, we have the same elements differently arranged. However, as the {356} word _tree-apple_ is not current in the language, the class of compounds indicated by it may seem to be merely imaginary. Nothing is farther from being the case. A _tree-rose_ is a rose of a particular sort. The generality of roses being on _shrubs_, this grows on a _tree_. Its peculiarity consists in this fact, and this particular character is expressed by the word _tree_ _prefixed_. A _rose-tree_ is a _tree_ of a particular sort, distinguished from _apple-trees_, and _trees_ in general (in other words, particularised or defined) by the word _rose_ _prefixed_.

A _ground-nut_ is a _nut_ particularised by growing in the ground. _A nut-ground_ is a _ground_ particularised by producing nuts.

A _finger-ring_, as distinguished from _ear-rings_, and from _rings_ in general (and so particularised), is a _ring_ for the _finger_. A _ring finger_, as distinguished from _fore-fingers_, and from _fingers_ in general (and so particularised), is a _finger_ whereon _rings_ are worn.

s. 412. At times this rule seems to be violated. The words _spitfire_ and _daredevil_ seem exceptions to it. At the first glance it seems, in the case of a _spitfire_, that what he (or she) _spits_ is _fire_; and that, in the case of a _daredevil_, what he (or she) _dares_ is the _devil_. In this case the initial words _spit_ and _dare_, are particularised by the final ones _fire_ and _devil_. The true idea, however, confirms the original rule. A _spitfire_ voids his fire by spitting. A _daredevil_, in meeting the fiend, would not shrink from him, but would defy him. A _spitfire_ is not one who spits fire, but one whose fire is _spit_. A _daredevil_ is not one who dares even the devil, but one by whom the devil is even dared.

s. 413. Of the two elements of a compound word, which is the most important? In one sense the latter, in another sense the former. The latter word is the most _essential_; since the general idea of _trees_ must exist before it can be defined or particularised; so becoming the idea which we have in _apple-tree_, _rose-tree_, &c. The former word, however, is the most _influential_. It is by this that the original idea is qualified. The latter word is the staple original element: the former is the superadded influencing element. Compared with each {357} other, the former element is active, the latter passive. Etymologically speaking, the former element, in English compounds, is the most important.

s. 414. Most numerous are the observations that bear upon the composition of words; _e.g._, how nouns combine with nouns, as in _sunbeam_; nouns with verbs, as in _daredevil_, &c. It is thought sufficient in the present work to be content with, 1. defining the meaning of the term composition; 2. explaining the nature of some obscure compounds.

Composition is the joining together, _in language_, of two _different words_, and _treating the combination as a single term_. Observe the words in italics.

_In language._--A great number of our compounds, like the word _merry-making_, are divided by the sign -, or the hyphen. It is very plain that if all words _spelt_ with a hyphen were to be considered as compounds, the formation of them would be not a matter of speech, or language, but one of writing or spelling. This distinguishes compounds in language from mere printers' compounds.

_Different._--In Old High German we find the form _selp-selpo_. Here there is the junction of two words, but not the junction of two _different_ ones. This distinguishes composition from gemination.--Grimm, Deutsche Grammatik, iii. 405.

_Words._--In _father-s_, _clear-er_, _four-th_, &c., there is the addition of a letter or a syllable, and it may be even of the part of a word. There is no addition, however, of a whole word. This distinguishes composition from derivation.

_Treating the combination as a single term._--In determining, in certain cases, between derived words and compound words, there is an occasional perplexity; the perplexity, however, is far greater in determining between a compound word and _two words_. In the eyes of one grammarian the term _mountain height_ may be as truly a compound word as _sunbeam_. In the eyes of another grammarian it may be no compound word, but two words, just as _Alpine height_ is two words; _mountain_ being dealt with as an adjective. It is in the determination of this that the accent plays an important part. This fact was foreshadowed in the Chapter upon Accents. {358}

s. 415. The attention of the reader is drawn to the following line, slightly altered, from Churchill:--

"Then r['e]st, my fri['e]nd, _and sp['a]re_ thy pr['e]cious br['e]ath."

On each of the syllables _rest_, _friend_, _spare_, _prec-_, _breath_, there is an accent. Each of these syllables must be compared with the one that precedes it; _rest_ with _then_, _friend_ with _my_, and so on throughout the line. Compared with the word _and_, the word _spare_ is not only accented, but the accent is conspicuous and prominent. There is so little on _and_, and so much on _spare_, that the disparity of accent is very manifest.

Now, if in the place of _and_, there was some other word, a word not so much accented as _spare_, but still more accented than _and_, this disparity would be diminished, and the accents of the two words might be said to be at _par_, or nearly so. As said before, the line was slightly altered from Churchill, the real reading being

Then r['e]st, my fri['e]nd, _spare, spare_ thy pr['e]cious breath.--

In the true reading we actually find what had previously only been supposed. In the words _spare, spare_, the accents are nearly at _par_. Such the difference between accent at _par_ and disparity of accent.

Good illustrations of the parity and disparity of accent may be drawn from certain names of places. Let there be such a sentence as the following: _the lime house near the bridge north of the new port._ Compare the parity of accent on the separate words _lime_ and _house_, _bridge_ and _north_, _new_ and _port_, with the disparity of accent in the compound words _L['i]mehouse_, _Br['i]dgenorth_, and _N['e]wport_. The separate words _beef steak_, where the accent is nearly at _par_, compared with the compound word _swe['e]pstakes_, where there is a great disparity of accent, are further illustrations of the same difference.

s. 416. The difference between a compound word and two words is greatest where the first is an adjective. This we see in comparing such terms as the following: _bl['a]ck b['i]rd_, meaning a _bird that is black_, with _bl['a]ckbird_=the Latin _merula_; or _bl['u]e b['e]ll_, meaning a _bell that is blue_, with _bl['u]ebell_, the flower. {359} Expressions like _a sh['a]rp edg['e]d instrument_, meaning _an instrument that is sharp and has edges_, as opposed to a _sh['a]rp-edged instrument_, meaning _an instrument with sharp edges_, further exemplify this difference.

Subject to four small classes of exceptions, it may be laid down, that, in the English language, _there is no composition unless there is either a change of form or a change of accent_.

The reader is now informed, that unless, in what has gone before, he has taken an exception to either a statement or an inference, he has either seen beyond what has been already laid down by the author, or else has read him with insufficient attention. This may be shown by drawing a distinction between a compound form and a compound idea.

In the words _a red house_, each word preserves its natural and original meaning, and the statement is _that a house is red_. By a parity of reasoning _a mad house_ should mean a _house that is mad_; and, provided that each word retain its natural meaning and its natural accent, such is the fact. Let a _house_ mean, as it often does, a _family_. Then the phrase, _a mad house_, means that the _house_, _or family_, _is mad_, just as a _red house_ means that the _house is red_. Such, however, is not the current meaning of the word. Every one knows that _a mad house_ means _a house for mad men_; in which case it is treated as a compound word, and has a marked accent on the first syllable, just as _L['i]mehouse_ has. Now, compared with the word _red house_, meaning a house of a _red colour_, and compared with the words _mad house_, meaning a _deranged family_, the word _m['a]dhouse_, in its common sense, expresses a compound idea; as opposed to two ideas, or a double idea. The word _beef steak_ is evidently a compound idea; but, as there is no disparity of accent, it is not a compound word. Its sense is compound; its form is not compound, but double. This indicates the objection anticipated, which is this: _viz._, that a definition, which would exclude such a word as _beef steak_ from the list of compounds, is, for that very reason, exceptionable. I answer to this, that the term in question is a compound idea, and not a compound form; in other words, that it is a compound in logic, but not a compound in etymology. {360} Now etymology, taking cognisance of forms only, has nothing to do with ideas, except so far as they influence forms.

Such is the commentary upon the words, "_treating the combination as a single term_;" in other words, such the difference between a compound word and two words. The rule, being repeated, stands (subject to the four classes of exceptions) thus: _There is no true composition without either a change of form or a change of accent._ As I wish to be clear upon this point, I shall illustrate the statement by its application.

The word _tr['e]e-rose_ is often pronounced _tr['e]e r['o]se_; that is, with the accent at _par_. It is compound in the one case; it is two words in the other.

The words _mountain ash_ and _mountain height_ are generally (perhaps always) pronounced with an equal accent on the syllables _mount-_ and _ash_, _mount-_ and _height_, respectively. In this case the word _mountain_ must be dealt with as an adjective, and the words considered as two. The word _mo['u]ntain wave_ is often pronounced with a visible diminution of accent on the last syllable. In this case there is a disparity of accent, and the word is compound.

s. 417. The following quotation indicates a further cause of perplexity in determining between compound words and two words:--

1.

A wet sheet and a blowing gale, A breeze that follows fast; That fills the white and swelling sail, And bends the _gallant mast_.

ALLAN CUNNINGHAM.

2.

Britannia needs no bulwarks, No towers along the steep; Her march is o'er the _mountain-wave_, Her home is on the deep.

THOMAS CAMPBELL.

To speak first of the word (or words) _gallant mast_. If _gallant_ mean _brave_, there are _two words_. If the words be two, there {361} is a stronger accent on _mast_. If the accent on _mast_ be stronger, the rhyme with _fast_ is more complete; in other words, the metre favours the notion of the words being considered as _two_. _Gallant-mast_, however, is a compound word, with an especial nautical meaning. In this case the accent is stronger on _gal-_ and weaker on _-mast_. This, however, is not the state of things that the metre favours. The same applies to _mountain wave_. The same person who in prose would throw a stronger accent on _mount-_ and a weaker one on _wave_ (so dealing with the word as a compound), might, in poetry, make the words _two_, by giving to the last syllable a parity of accent.

The following quotation from Ben Jonson may be read in two ways; and the accent may vary with the reading.

1.

Lay thy bow of pearl apart, And thy _silver shining_ quiver.

2.

Lay thy bow of pearl apart, And thy _silver-shining_ quiver.

_Cynthia's Revels._

s. 418. _On certain words wherein the fact of their being compound is obscured._--Composition is the addition of a word to a word, derivation is the addition of letters or syllables to a word. In a compound form each element has a separate and independent existence; in a derived form, only one of the elements has such. Now it is very possible that in an older stage of a language two words may exist, may be put together, and may so form a compound; at the time in point each word having a separate and independent existence: whilst, in a later stage of language, only one of these words may have a separate and independent existence, the other having become obsolete. In this case a compound word would take the appearance of a derived one, since but one of its elements could be exhibited as a separate and independent word. Such is the case with, amongst others, the word _bishopric_. In the present language the word _ric_ has no separate and independent existence. For all this, the word {362} is a true compound, since, in Anglo-Saxon, we have the noun _r['i]ce_ as a separate, independent word, signifying _kingdom_ or domain.

Again, without becoming obsolete, a word may alter its form. This is the case with most of our adjectives in _-ly_. At present they appear derivative; their termination _-ly_ having no separate and independent existence. The older language, however, shows that they are compounds; since _-ly_ is nothing else than _-lic_, Anglo-Saxon; _-lih_, Old High German; _-leiks_, Moeso-Gothic;=_like_, or _similis_, and equally with it an independent separate word.

For the following words a separate independent root is presumed rather than shown. It is presumed, however, on grounds that satisfy the etymologist.

_Mis-_, as in _misdeed_, &c.--Moeso-Gothic, _miss[^o]_=_in turns_; Old Norse, _[^a] mis_=_alternately_; Middle High German, _misse_=_mistake_. The original notion _alternation_, thence _change_, thence _defect_. Compare the Greek [Greek: allos].--Grimm, Deutsche Grammatik, ii. 470.

_Dom_, as in _wisdom_, &c.--The substantive _d[^o]m_ presumed.--Deutsche Grammatik, ii. 491.

_Hood_ and _head_, as in _Godhead_, _manhood_, &c.--The substantive _h['a]ids_=_person_, _order_, _kind_, presumed.--Deutsche Grammatik, ii. 497. Nothing to do with the word _head_.

_Ship_, as in _friendship_.--Anglo-Saxon, _-scipe_ and _-sceaeft_; German, _-schaft_; Moeso-Gothic, _gaskafts_=_a creature_, or _creation_. The substantive _skafts_ or _skap_ presumed. The _-skip_ or _-scape_ in _landskip_ is only an older form.--Deutsche Grammatik, ii. 522.

_Less_, as in _sleepless_, &c., has nothing to do with _less_. Derived from _l['a]us_, _l[^o]s_, _destitute of_=Latin, _expers_.--Deutsche Grammatik, ii. 565.

For the further details, which are very numerous, see the Deutsche Grammatik, vol. iii.

s. 419. "Subject to four classes of exceptions, it may be laid down that _there is no true composition unless there is either a change of form or a change of accent_."--Such is the statement made in p. 359. The first class of exceptions consists {363} of those words where the natural tendency to disparity of accent is traversed by some rule of euphony. For example, let two words be put together, which at their point of contact form a combination of sounds foreign to our habits of pronunciation. The rarity of the combination will cause an effort in utterance. The effort in utterance will cause an accent to be laid on the latter half of the compound. This will equalize the accent, and abolish the disparity. The word _monkshood_, the name of a flower (_aconitum napellus_), where, to my ear at least, there is quite as much accent on the _-hood_ as on the _monks-_, may serve in the way of illustration. Monks is one word, hood another. When joined together, the _h-_ of the _-hood_ is put in immediate opposition with the _-s_ of the _monks-_. Hence the combination _monkshood_. At the letters _s_ and _h_ is the point of contact. Now the sound of _s_ followed immediately by the sound of _h_ is a true aspirate. But true aspirates are rare in the English language. Being of rare occurrence, the pronunciation of them is a matter of attention and effort; and this attention and effort creates an accent which otherwise would be absent. Hence words like _monksh['o]od_, _well-h['e]ad_, and some others.

Real reduplications of consonants, as in _hop-pole_, may have the same parity of accent with the true aspirates: and for the same reasons. They are rare combinations that require effort and attention.

The second class of exceptions contains those words wherein between the first element and the second there is so great a disparity, either in the length of the vowel, or the length of the syllable _en masse_, as to counteract the natural tendency of the first element to become accented. One of the few specimens of this class (which after all may consist of double words) is the term _upst['a]nding_. Here it should be remembered, that words like _haph['a]zard_, _foolh['a]rdy_, _uph['o]lder_, and _withh['o]ld_ come under the first class of the exceptions.

The third class of exceptions contains words like _perch['a]nce_ and _perh['a]ps_. In all respects but one these are double words, just as _by chance_ is a double word. _Per_, however, differs from _by_ in having no separate existence. This sort of words {364} we owe to the multiplicity of elements (classical and Gothic) in the English language.

To anticipate objections to the rule respecting the disparity of accent, it may be well to state in fresh terms a fact already indicated, viz., that the same combination of words may in one sense be compound, and in the other double (or two). _An uphill game_ gives us the combination _up_ + _hill_ as a compound. _He ran up hill_ gives us the combination _up_ + _hill_ as two words. So it is with _down_ + _hill_, _down_ + _right_, and other words. _Man-servant_, _cock-sparrow_, &c., are double or compound, as they are pronounced _m['a]n-s['e]rvant_, _m['a]n-servant_, _c['o]ck-sp['a]rrow_, or _c['o]ck-sparrow_.

The fourth class is hypothetical. I can, however, imagine that certain compounds may, if used almost exclusively in poetry, and with the accent at _par_, become so accented even in the current language.

s. 420. For a remark on the words _peacock_, _peahen_, see the Chapter upon Gender.--If these words be rendered masculine or feminine by the addition of the elements _-cock_ and _-hen_, the statements made in the beginning of the present chapter are invalidated. Since, if the word _pea-_ be particularized, qualified, or defined by the words _-cock_ and _-hen_, the second term defines or particularises the first, which is contrary to the rule of p. 355. The truth, however, is, that the words _-cock_ and _-hen_ are defined by the prefix _pea-_. Preparatory to the exhibition of this, let us remember that the word _pea_ (although now found in composition only) is a true and independent substantive, the name of a species of fowl, like _pheasant_, _partridge_, or any other appellation. It is the Latin _pavo_, German _pfau_. Now, if the word _peacock_ mean a _pea_ (_pfau_ or _pavo_) that is a male, then do _wood-cock_, _black-cock_, and _bantam-cock_, mean _woods_, _blacks_, and _bantams_ that are male. Or if the word _peahen_ mean a _pea_ (_pfau_ or _pavo_) that is female, then do _moorhen_ and _guineahen_ mean _moors_ and _guineas_ that are female. Again, if a _peahen_ mean a _pea_ (_pfau_ or _pavo_) that is female, then does the compound _pheasant-hen_ mean the same as _hen-pheasant_; which is not the case. The fact is that _peacock_ means a _cock that is a pea_ (_pfau_ or _pavo_); {365} _peahen_ means a _hen that is a pea_ (_pfau_ or _pavo_); and, finally, _peafowl_ means a _fowl that is a pea_ (_pfau_ or _pavo_). In the same way _moorfowl_ means, not a _moor that is connected with a fowl_, but a _fowl that is connected with a moor_.

s. 421. It must be clear, _ex vi termini_, that in every compound word there are two parts; _i. e._, the whole or part of the original, and the whole or part of the superadded word. In the most perfect forms of inflection there is a third element, _viz._, a vowel, consonant, or syllable that joins the first word with the second.

In the older forms of all the Gothic languages the presence of this third element was the rule rather than the exception. In the present English it exists in but few words.

_a._ The _-a-_ in _black-a-moor_ is possibly such a connecting element.

_b._ The _-in-_ in _night-in-gale_ is most probably such a connecting element. Compare the German form _nacht-i-gale_, and remember the tendency of vowels to take the sound of _-ng_ before _g_.

s. 422. _Improper compounds._--The _-s-_ in words like _Thur-s-day_, _hunt-s-man_, may be one of two things.

_a._ It may be the sign of the genitive case, so that _Thursday_=_Thoris dies_. In this case the word is an improper compound, since it is like the word _pater-familias_ in Latin, in a common state of syntactical construction.

_b._ It may be a connecting sound, like the _-i-_ in _nacht-i-gale_. Reasons for this view occur in the following fact:--

In the Modern German languages the genitive case of feminine nouns ends otherwise than in _-s_. Nevertheless, the sound of _-s-_ occurs in composition equally, whether the noun it follows be masculine or feminine. This fact, as far as it goes, makes it convenient to consider the sound in question as a connective rather than a case. Probably, it is neither one nor the other exactly, but the effect of a false analogy.

s. 423. _Decomposites._--"Composition is the joining together of _two_ words."--See p. 357.

In the first edition the sentence ran "_two or more_" words; being so written to account for compounds like _mid-ship-man_, {366} _gentle-man-like_, &c., where the number of verbal elements seems to amount to three.

Nevertheless, the caution was unnecessary. Compound radicals like _midship_ and _gentleman_, are, for the purposes of composition, single words. Compounds wherein one element is compound are called decomposites.

s. 424. The present chapter closes with the notice of two classes of words. They are mentioned now, not because they are compounds, but because they can be treated of here more conveniently than elsewhere.

There are a number of words which are never found by themselves; or, if so found, have never the same sense that they have in combination. Mark the word combination. The terms in question are points of combination, not of composition: since they form not the parts of words, but the parts of phrases. Such are the expressions _time and tide_--_might and main_--_rede me my riddle_--_pay your shot_--_rhyme and reason_, &c. These words are evidently of the same class, though not of the same species with _bishopric_, _colewort_, _spillikin_, _gossip_, _mainswearer_, and the words quoted in p. 362. These last-mentioned terms give us obsolete words preserved in composition. The former give us obsolete words preserved in combination.

The other words are etymological curiosities. They may occur in any language. The English, however, from the extent of its classical element, is particularly abundant in them. It is a mere accident that they are all compound words.

* * * * *

{367}