Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, LINGUISTICS ( (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 23 February 2017


Summary and Keywords

Blocking can be defined as the non-occurrence of some linguistic form, whose existence could be expected on general grounds, due to the existence of a rival form. *Oxes, for example, is blocked by oxen, *stealer by thief. Although blocking is closely associated with morphology, in reality the competing “forms” can not only be morphemes or words, but can also be syntactic units. In German, for example, the compound Rotwein ‘red wine’ blocks the phrasal unit *roter Wein (in the relevant sense), just as the phrasal unit rote Rübe ‘beetroot; lit. red beet’ blocks the compound *Rotrübe. In these examples, one crucial factor determining blocking is synonymy; speakers apparently have a deep-rooted presumption against synonyms. Whether homonymy can also lead to a similar avoidance strategy, is still controversial. But even if homonymy blocking exists, it certainly is much less systematic than synonymy blocking.

In all the examples mentioned above, it is a word stored in the mental lexicon that blocks a rival formation. However, besides such cases of lexical blocking, one can observe blocking among productive patterns. Dutch has three suffixes for deriving agent nouns from verbal bases, -er, -der, and -aar. Of these three suffixes, the first one is the default choice, while -der and -aar are chosen in very specific phonological environments: as Geert Booij describes in The Morphology of Dutch (2002), “the suffix -aar occurs after stems ending in a coronal sonorant consonant preceded by schwa, and -der occurs after stems ending in /r/” (p. 122). Contrary to lexical blocking, the effect of this kind of pattern blocking does not depend on words stored in the mental lexicon and their token frequency but on abstract features (in the case at hand, phonological features).

Blocking was first recognized by the Indian grammarian Pāṇini in the 5th or 4th century bc, when he stated that of two competing rules, the more restricted one had precedence. In the 1960s, this insight was revived by generative grammarians under the name “Elsewhere Principle,” which is still used in several grammatical theories (Distributed Morphology and Paradigm Function Morphology, among others). Alternatively, other theories, which go back to the German linguist Hermann Paul, have tackled the phenomenon on the basis of the mental lexicon. The great advantage of this latter approach is that it can account, in a natural way, for the crucial role played by frequency. Frequency is also crucial in the most promising theory, so-called statistical pre-emption, of how blocking can be learned.

Keywords: blocking, pre-emption, suppletion, competition, Elsewhere Condition, Pāṇini’s principle, synonymy, homonymy

1. Historical and Terminological Issues

The English term blocking was launched in Aronoff (1976), where it was defined as “the non-occurrence of one form due to the simple existence of another” (p. 43). In the context of Aronoff’s book, the concept of blocking explained the complementary distribution of abstract nouns such as those of the second and third columns under (1):


















The oddness of *gloriosity or *furiosity was attributed to their being blocked by the synonymous existing words glory and fury. The fact that gloriousness and furiousness do not share the same fate was said to be due to the unlimited productivity of the suffix -ness. This unlimited productivity was believed to entail that derivatives with this suffix need not be entered into the permanent lexicon, since “[t]he most productive classes never have to be listed” (p. 45). Aronoff conceived of the lexicon as “arranged according to stems,” with “a slot for each canonical meaning” capable of housing only one item (p. 45). If a complex word wanted to lodge itself in one of these slots of the permanent lexicon, but the slot was already occupied by another host, the newcomer was turned away.

The above treatment was the first, though only partially successful, attempt in modern linguistics to provide an account of the phenomenon in the framework of an explicit theory of the lexicon, but it was by no means the first time that the phenomenon itself had been subjected to scrutiny by scholars. The Indian grammarian Pāṇini (5th or 4th century bc) is generally credited as the first one to treat the phenomenon of blocking. In his description of Sanskrit, he followed a principle according to which, of two rules applicable to the same item, the more specific one had precedence, a principle first formulated explicitly by the later grammarian Patañjali (c. 150 bc; Janda & Sandoval, 1984, pp. 2−3). Some linguists call the principle “Pāṇini’s principle” in honor of its inventor.

Towards the end of the 19th century, the phenomenon of blocking was given explicit attention by two heavy weights of German linguistics, namely, Hermann Paul and Hermann Osthoff. In his classic article from 1896, Paul articulated the general insight that “the competition between synonymous means of expression is an essential factor in the historical development [sc. of languages]” (p. 19). At a more concrete level, he showed that “the formation of [German action] nouns in -ung was prevented [verhindert] by the existence of simpler formations with the function of an action noun” (p. 27). Cases in point, according to Paul (p. 30), are Lob ‘praise’, Raub ‘theft’, Wahl ‘choice’, Verlust ‘loss’, Gebrauch ‘use’, Verbrauch ‘consumption’,” Verkauf ‘sale’, or Ankauf ‘purchase’. All these nouns can govern a direct object and prevent the coining of a corresponding noun with the productive suffix -ung: das Lob/*die Lobung der Torheit “the praise of foolishness,” etc. Paul also correctly stated that “it is not excluded that synonyms may arise and stay in the language side by side for a long time” (p. 34). Last but not least, he observed that such “overabundance” [Überfluss] had repeatedly been put to use for semantic differentiation (p. 35). After Paul’s article, many authors treated blocking in his style, at least in the German-speaking countries, but no further theoretical progress was made. The only noteworthy event was the introduction of a new German terminology by Schindler (1972, p. 43), who referred to the phenomenon as Blockierung.

Three years after Paul, in 1899, Osthoff published the foundational essay on what has ever since been called “suppletion” (Suppletivwesen, Suppletivismus; cf. his claim to authorship concerning this terminology on p. 3). In Osthoff (1899), suppletion refers to the filling of one or several slots of a paradigm with a formally unrelated substitute instead of the regularly derived form to be expected on general grounds, as when the comparative slot of good is filled with better instead of *gooder. The relationship between better and the paradigm of good is interpreted as one of “complementing” [Ergänzung] (in fact, Ergänzungswesen is used throughout the paper as a synonym of Suppletivwesen). Osthoff (pp. 4−5) explicitly rejected an interpretation of words such as good as “defective.” It is not the case, according to him, that better moves into a slot that good would, in principle, be unable to fill. Rather, he stresses the collaborative nature of the relationship, speaking of “a substitution, a mutual helping out and complementing each other” [eine stellvertretung, ein gegenseitiges sichaushelfen und sichergänzen] (p. 4). In this conception, the paradigm is viewed as a whole, while those who use the blocking terminology focus on the relationship between the two contenders for a slot, treating the blocking word as a sort of squatter hindering the legitimate owner from entering his home. In view of our close reading of Osthoff’s seminal work, the critique of the term suppletion, in Levin (1972), does not seem to be entirely justified, at least with respect to Osthoff himself. Levin considered suppletion to be a misleading term if taken in its etymological sense of “filling in a gap,” since there is no gap in the first place. However, as we have seen, Osthoff also explicitly rejected the defectivity view. Levin proposed to talk of pre-emption instead, a term that has also gained ground in the literature written in English as a synonym of blocking.

Before we turn to a more thorough discussion of some theoretical issues related to blocking, it may be useful to give a survey of the range of phenomena that have been, or should be, subsumed under this notion.

2. The Range of Phenomena Subsumed Under Blocking

The prototypical cases used for illustrating the phenomenon of blocking are generally drawn from the realm of morphology, both word formation and inflection.

In inflection, the most frequently cited examples are irregular forms such as oxen (blocking *oxes), better (blocking *gooder), or went (blocking *goed). In general, things are presented this way, as blocking of one word form by another. Other scholars instead speak of a blocking effect exerted by stems or affixes on rival stems or affixes. In that case, one would say that the plural suffix -en blocks -es with the base ox, or that the stem bett- blocks good in the context of the comparative suffix -er. In the case of went, the blocking word is probably most naturally considered as simplex, though decomposition into a stem wen- and a suffix -t (cf. buil-t, and similar past-tense forms) is not completely out of the question. In inflection, the blocking effect is often absolute, at least in the case of adult speakers and under normal circumstances. Nevertheless, there are also cases where two or more forms compete for one and the same slot (e.g., dreamed and dreamt) even in the idiolect of one and the same speaker, a phenomenon that is currently discussed under the heading of “overabundance” (Thornton, 2012).

In research on inflection, it is more common to speak of “suppletion” than of “blocking,” because the discussion generally revolves around paradigm organization and not around the relationship between an irregular form and the corresponding suppressed regular form. A further difference between the use of the terms blocking and suppletion consists in the fact that only the latter terminology requires the blocking word form, stem, or affix to be formally unrelated to the form that is blocked. If there is no formal relationship at all, as in went/*goed, it is customary to speak of “full suppletion,” while in “partial suppletion,” there still is some formal overlap, while the divergence is not derivable by a phonological rule. A case in point would be the corresponding participle gone, which conserves some phonological resemblance to go (more obvious to the eye than to the ear, for sure).

Inflectional classes are not normally discussed under the heading of “blocking,” but it is not difficult to see the kinship with the English cases adduced above. In Latin, for example, stems referring to males go into the -us/-i class by default (e.g., dominus, -i ‘master’; servus, -i ‘slave’). A sizeable number of stems, however, like naut- ‘sailor’, take the ending of the -a/-ae class, which is productive with nouns denoting females, but unproductive with nouns referring to males. The inflectional behavior of each of these latter nouns must therefore be learned individually. We can assume that each noun denoting a male that follows the -a/-ae class blocks the corresponding word in -us/-i. Evidence for such an assumption is provided by occasional defections of males in -a/-ae towards the -us/-i class, if the blocking force withers for some reason, as in Late Latin piratus, -i instead of pirata, -ae (cf., Gardani, 2013, p. 210). The case is entirely parallel to oxen/*oxes, apart from the fact that the class of words affected is much bigger and therefore often referred to in grammars as a “class,” while oxen is simply listed as an isolated exception.

In the case just discussed, the set of male bases adopting -a/-ae is lexically governed. In other cases, by contrast, the blocking words belong to an open class definable by means of some common feature. A case in point is constituted by the realization of the second singular indefinite present indicative in Hungarian (Carstairs, 1988, p. 70): -(a)sz is the default ending, but after sibilants and affricates -ol must be used instead. The verb stems that require -ol need not be enumerated individually; even neologisms that happen to end in sibilants or affricates would automatically take -ol instead of -(a)sz. The blocking relationship in cases like this one must therefore be defined not at the level of individual items, but at the more abstract level of patterns (constructions, rules, or whatever terminology is used).

As we have seen in section 1, “Historical and Terminological Issues,” Aronoff introduced the term blocking in the context of derivational morphology. The following cases provide an idea of the broad range of phenomena that have been, or should be, subsumed under “blocking” in this area.

Let us start with the German action nouns already mentioned in section 1. As pointed out by Paul, the productive suffix -ung is occasionally blocked by well-entrenched shorter synonyms such as Lob ‘praise’, etc. The set of blocking words forms a motley collection with no common denominator, neither formal nor semantic. Each of these words blocks its hypothetical rival in -ung on an individual basis. Cases of this kind, where one or a few formations of a productive series are blocked by existing synonyms, are legion. A staple case for English is the absence of the expected agent noun *cooker due to the existence of cook. Note, however, that *cooker is only blocked in sense of “cook,” not in the instrumental sense of “appliance for cooking,” which is not covered by cook, nor in the more specialized agentive sense “person employed to operate cooking apparatuses in the commercial preparation of food and drink.” Since Kiparsky (1983, p. 9), constellations of this kind have been referred to as “partial blocking.” In both the German case and the case of cook/*cooker, the blocking word is formed according to a synchronically unproductive pattern of conversion. In other cases, such as thief/*stealer, the blocking word is a suppletive stem, very much like in better/*gooder or went/*goed. However, as the following examples show, the blocking word can also be itself the output of a productive rule.

Miyagawa (1984) argued that the two Japanese causative suffixes -sas and -sase were both fully productive, but that with single verbs lexical blocking by an entrenched rival formation could be observed. In a similar vein, Polish, according to Malicka-Kleparska (1985), has two productive suffixes for forming diminutives from masculine nouns, -ek and -ik (both with allomorphic variants that need not concern us here). In principle, both these suffixes can attach to any masculine noun. The actual distribution, however, is different: while with the majority of bases the two suffixes are indeed used side by side, in some cases one of the two suffixes is so well entrenched that it effectively blocks the rival suffix. Under (2), illustrative examples are displayed for the three distributional classes: the bases under a admit both suffixes, those under b have a preference for -ek, and those under c, for -ik.



łom ‘crowbar’



gnat ‘bone’




dom ‘house’


kwiat ‘flower’



tom ‘volume’


temat ‘theme’


A similar case from English is represented by the rivalry of the adjectival suffixes -ic and -ical. As shown in Lindsay and Aronoff (2013, pp. 141−148), both suffixes have been productive since Middle English, but “[i]n general, most stems clearly favored one suffix over the other” (p. 143). While Olympia, polyphony, and sulphur prefer -ic to the point of exclusivity, mathematics, surgery, and theology have an equally clear predilection for -ical. Overall, -ic is preferred at a ratio of roughly 10 to 1, but -ical has managed to become the favorite suffix in the case of the large group of nouns in -ology. This case therefore combines the Japanese/Polish scenario of competition at the level of individual words formed with productive suffixes and the specialization of one of the suffixes in a niche that can be defined at a more general morphosemantic level (bases in -ology).

Such cases, where one general pattern is curtailed by a more specific one, are quite common in word formation. The case of Dutch agent nouns in -er, -der, and -aar has already been mentioned in the “Summary” at the beginning of this essay. A similar case from English is constituted by the suffixes -ify and -ize, which show a complementary distribution determined by prosodic factors (cf. Plag, 1999, pp. 196−197): -ify requires the preceding syllable to carry main stress (e.g., fálsify, opácify), while -ize needs trochaic or dactylic bases (e.g., rándomize, colónialize). In this English example, the default status of -ize is less obvious than the default status of agentive -er in Dutch. However, a set-subset relation, or at least a partial overlap of the domains is crucial in order to conceive of the competition as a case of blocking.

Up to now, we have seen cases of blocking among word forms in inflection and among lexemes in word formation, either at the level of individual items or at a more general level definable in terms of semantic or formal features. It has often been observed, however, that blocking is not confined to the morphology of a language. The English comparative is normally summoned as witness for this claim. The details of comparative formation in English are quite intricate, and much ink has been spilt recently on this subject (cf. Adams 2012; González-Díaz, 2008; Mondorf, 2009). In the present context, it is sufficient to say that many adjectives reject the comparative suffix -er, for example, trisyllabic adjectives such as curious or extensive. In these cases, the comparative is formed analytically by means of the adverb more: more curious/*curiouser, more extensive/*extensiver, etc. This complementary distribution is often interpreted as a case of blocking exerted on the syntactic pattern with more by the morphological comparatives; that is, blocking here transcends the morphology-syntax boundary. Syntax can also suffer blocking from lexical items which, from an historical perspective, are often amalgamated forms, such as French du and au, which block the regular syntactic sequences de le “of the” and à le “to the.” While in this case the blocking is categorical, in other cases the full form and the amalgamated form continue to be used side by side, as in German am and auf dem ‘on the’.

In the preceding cases, morphology or the lexicon block syntax. The reverse can also be observed, but only if the syntactic pattern has been co-opted by the lexicon as a naming device. German, for example, uses both phrasal units consisting of adjective + noun and adjective-noun compounds for naming: the concept “red wine” is expressed by the compound Rotwein, the concept “beetroot,” by contrast, by the phrase rote Rübe, literally “red beet” (the -e of rote is a feminine ending marking agreement with the feminine noun Rübe). Both the compound and the phrase normally block the alternative form (cf. *roter Wein, *Rotrübe, in the relevant sense). In their study of the rivalry between these formations, Schlücker and Plag (2011, p. 1540) observe: “As syntax normally does not block word formation, the blocking ability of these phrases indicates their lexical status.”

The following example shows that blocking can also obtain among syntactic patterns. In Dominican Spanish, two presentational patterns compete with one another, <AdvP haber Obj> (e.g., ha habido ciertos cambios ‘there have been certain changes’, with 3rd person singular ha) and <AdvP haber Subj> (e.g., han habido ciertos cambios, where han agrees in number with the plural subject cambios). The first pattern is the older one, but the second, more colloquial pattern is increasingly encroaching on its territory. As Claes (2015) has shown, the blocking force of the older pattern is not evenly distributed over the different tenses, but depends on their frequency: “the tenses for which an entrenched instance of the singular variant was posited (i.e., present and preterit) disfavour the use of the pluralized presentational haber construction strongly […]. In contrast, all other types of expressions are more frequently formed with pluralized presentational haber” (p. 19). As one can see, the blocking effect is not located here at the level of the patterns as such, but at the level of specific tensed forms. Furthermore, the effect is not one of all-or-none, but a question of higher or lower probability of occurrence.

Last but not least, blocking can also take place between the lexicon or word formation and patterns of semantic extension. Brdar (2009, pp. 65−72), for example, argues that the metonymic extension of the type “animal” > “meat of the animal” (e.g., I would not eat cat) is blocked in English with the nouns cow, calf, pig, sheep, and deer because of the existence of the well-entrenched special designations beef, veal, pork, mutton, and venison. In all but one of these cases, French, the donor language of the English meat terms, has regular metonymy (boeuf, veau, porc, mouton; venaison ‘venison’ is flanked by gibier ‘game’). In the Germanic languages, compounds of the type German Schaffleisch, literally ‘sheep meat’, tend to be preferred to metonymic extension, and in Slavic languages the same is true for suffixations in -ina of the type Croatian govedina ‘beef’. As the author shows, however, the interplay of the lexicon, word formation, and semantic extension can be quite intricate; the cut-off point between the patterns is somewhat different in each language, and the strength of the blocking force is also subject to variation.

3. Theoretical Issues

3.1. Synonymy vs. Homonymy Blocking

Our preliminary definition of blocking in the introduction as “the non-occurrence of one form due to the simple existence of another” (Aronoff, 1976, p. 43) does not specify what it is exactly that causes the blocking. Two factors have been put forward in the literature, synonymy and homonymy.

All the cases discussed in Section 2, “The Range of Phenomena Subsumed Under Blocking,” were cases of synonymy blocking. Suppletion in inflection (e.g., went/*goed) and derivation (e.g., thief/*stealer) as well as the phenomenon of partial blocking leave no doubt about the fact that synonymy is a crucial factor in many cases. It could also be the case that synonymy (or lexicalization, which amounts to the same) plays a role in explaining why the plural mice does not completely block mouses in computer mouses (134,000 hits on Google on October 22, 2015, as opposed to 508,000 for computer mice), and similar cases. At the same time, it is clear that synonymy is not a sufficient condition, as we will see in Section 3.3, “The Relativity of Blocking.” Whether it should be considered as a necessary condition depends on the stance one takes with respect to homonymy blocking.

There seems to be a general consensus concerning the existence of a deep presumption of speakers against synonymy. As early as 1897, Michel Bréal elevated it to the status of a “law” (“la loi de répartition,” Bréal, 1897, Chapter 2). A question about which less consensus seems to exist is why this presumption against synonymy exits. For Bréal, it was an intuitive feeling of speakers that different words must mean different things. Similarly, for Gauger (1972, p. 34) it was a corollary of the naming function of words. As we have already seen in Section 1, “Historical and Terminological Issues,” Aronoff (1976) attributed it to the organization of the lexicon in slots that would allow at most one permanent host. McCawley (1978, p. 246), two years later, resorted to the principle of least effort in order to explain why pink blocks pale red: “pale red can be said to involve more effort than pink.” Pink therefore “is to be preferred over pale red unless something demands the extra effort involved in saying the latter.” Due to this general rule, pale red, according to McCawley, if used, is interpreted by conversational implicature as referring to a color distinct from pink, closer to red than pink. Kiparsky (1983, p. 16) surmised that his “Avoid Synonymy” principle might turn out to be better viewed “as a language learning device rather than as a formal constraint of grammar.” Other authors assume that the presumption against synonymy is somehow part of our genetic endowment as human beings: Clark (1993, p. 91) showed that a “principle of contrast” leading children to assume that “differences in form mark differences in meaning” is already operative early on in language acquisition, while Carstairs-McCarthy (2010, p. 59) attributed deep phylogenetic roots to the presumption against synonymy, which he argues to be shared by apes and even “one dog.” In Lindsay and Aronoff’s (2013) view of glossogenetic evolution, intolerance of true synonymy is also taken as a given. According to these authors, of two competing synonymous patterns, the less productive one can only survive in the long run if it firmly establishes itself in a clearly defined niche.

A side issue, vaguely related to the question of synonymy, is that of hypercharacterization. Blocking should not only prevent *oxes instead of oxen, but also *oxens, that is, the attachment of a second affix with the same meaning, except, of course, where multiple affixation makes sense (e.g., in anti-anti-aircraft missile). Multiple affixation of this kind, however, is quite frequent in the languages of the world (cf. Gardani, 2015). A notorious case from derivational morphology is German Prinzessin ‘princess’, coined in the 17th century by attaching the productive feminizing suffix -in to Prinzeß ‘princess’ (instead of the masculine base Prinz ‘prince’), which contained the unproductive feminizing suffix - taken over from French (-esse). Why didn’t speakers prefer the regular formation Prinzin to Prinzessin? It is possible that Prinzeß had become somewhat opaque as far as its internal structure was concerned, mitigating in that way the pleonastic effect. But overall cases of hypercharacterization are too varied to be amenable to a monocausal explanation.

Much more controversial than synonymy blocking is the existence of blocking by a homonym. It has been argued, for example, that the inexistence of the meaning “to pass the X” for spring and fall, as opposed to summer and winter (e.g., to summer/winter in California vs. *to spring/fall in California), was due to the existence of spring and fall as verbs of movement. The difficulty with such anecdotal evidence is that alternative explanations are available: summer and winter, for example, are typical vacation seasons, but not spring and fall. In support of this latter explanation, one could point to the fact that autumn, though not blocked by a homonymous verb, does not seem to fare much better in the sense “to pass the X” than fall. Large-scale systematic studies of homonymy blocking, unfortunately, are scarce. The most thorough treatment I know of can be found in Plank (1981, pp. 165−173), who could not find any systematic blocking effect in German. Carstairs-McCarthy (2010, p. 212) also concludes that “[t]here seems to be no general cognitively based presumption against homonymy among words, comparable to the presumption against synonymy.”

3.2. Lexical Blocking vs. Pattern Blocking

In Aronoff (1976), the term blocking was applied not only to the competition between an established word form or lexeme and a rival word form or lexeme, but also to cases of competition between two (or more) productive patterns: “An affix which is productive with a given morphological class will […] block the attachment of rival affixes to that class” (p. 45). In the discussion of blocking during the 1980s (cf. Plank, 1981, p. 180; Rainer, 1989), however, it soon became apparent that these two kinds of competition should better be kept apart, since they obey partly different conditions, a fact also acknowledged in Aronoff (1994). The two kinds of blocking have received different names in the literature; in the present article, they are called “lexical blocking” and “pattern blocking” respectively.

What are the properties common to both phenomena, and where lie the differences? Both of them crucially involve synonymy and competition. Both of them also involve at least one productive pattern (but cf. Section 3.4, “Actual, Possible, and Potential Words” on the dissonant view of distributed morphologists). The main difference lies in the fact that in one case, the blocking element is a lexical item—word form, lexeme or phrasal unit—stored in the mental lexicon, while in the case of two competing productive patterns, the blocking pattern need only be defined by way of an intensional definition: pattern A takes precedence over pattern B whenever a certain base fits the crucial requirements of A better than those of B. In principle, the differentiation between lexical blocking and pattern blocking is clear-cut. Nevertheless, as the English suffixes -ic and -ical presented in Section 2, “The Range of Phenomena Subsumed Under Blocking” have shown, lexical blocking and pattern blocking are often intimately intertwined.

3.3. The Relativity of Blocking

Blocking is not a question of all or none, neither lexical blocking nor pattern blocking. Rather, the actual blocking force depends on a series of factors that partly differ in lexical blocking and pattern blocking.

Paul (1920, pp. 114−116) provided a psycholinguistic account of lexical blocking as competition between a stored item and an “analogy,” a productive pattern. He also established a positive correlation between the token frequency of the blocking item and its blocking force, its resistance to regularization; the more frequent the blocking word, the higher the probability that it will win over the potential rival formed according to a productive pattern (cf. p. 208). In the light of this hypothesis, the fact that regularizations are more common in child language than in the language of adults was interpreted by him as a consequence of their memory traces still being weaker. Paul’s hypothesis also nicely accounts for the observation that lexical blocking seems to be stricter in inflection than in word formation, due to the higher token frequency of irregular inflectional forms. Kiparsky (2010, p. 319), however, has pointed out that the tighter paradigmatic organization of inflection, which guarantees strict synonymy, could also explain part of the difference. Another factor, hinted at in Rainer (1989, p. 164), is the degree of productivity of the rival pattern: the blocking force is interpreted “as the result of the antagonism between the pressure exerted by a potential regular word and the resistance offered by the corresponding blocking word, whereby pressure is a function of productivity and resistance a function of frequency.” Statistical evidence for this view has been provided by Maslen, Theakston, Lieven, and Tomase (2004), whose dense-corpus study of overgeneralizations in child speech has shown that these cannot be attributed solely to failure of retrieval of low-frequency forms, as Paul had thought, but depend also on the strength of the regular pattern involved.

The blocking force is also highly variable in pattern blocking, as already observed in Section 2 “The Range of Phenomena Subsumed Under Blocking.” In some cases, pattern A neatly curtails the domain of pattern B; the Dutch suffixes -der and -aar, which restrict the domain of -er, would be cases in point. At the other end of the scale, we encountered cases of more or less free variation, such as the two Japanese causative suffixes or the two Polish diminutive suffixes, which are said to co-occur freely as long as no lexical blocking is involved. Detailed studies of pattern blocking are still few and far between, which is why it would be premature to make any sweeping statements. Among the factors that may be expected to influence the blocking force one might surmise the neatness of the definition of the boundaries of the blocking pattern to be of central importance. In the Dutch case mentioned above, for example, the stem-final phonemes provide a clear criterion for selecting -der and -aar over -er. Such clear-cut delimitations of the domain of a pattern, however, are rather the exception than the rule in word formation, and even a pattern with a clear-cut domain need not necessarily block a more general rival pattern. In Spanish, for example, the relational suffix -eano is always possible with proper names ending in -e (e.g., Sartresartreano), but it does not hinder the attachment of the default suffix -iano (cf. sartriano). The two patterns are in free variation. Very often in word formation the attachment of an affix is conditioned not by one neat criterion but by a series of factors, each of which enhances or diminishes the probability of the affix being selected. As a consequence, the choice of A over B, or B over A, is determined by weighing negative probabilities against positive ones, a situation most adequately handled by some variant of analogical modelling (cf. Schlücker & Plag, 2011).

Whether blocking or free variation obtains between two competing patterns has to be learned from the data and is not deducible from a typology of domains, as proposed in van Marle’s (1986) “domain hypothesis.” But how can speakers learn to not use a certain pattern despite its semantic fit? The most promising answer to this crucial question seems to be that speakers do so by “statistical pre-emption.” According to Goldberg (2011), the strength of statistical pre-emption is determined by the probability with which a certain item occurs in pattern A in a context in which both pattern A and B should be possible in principle, as well as by the absolute frequency with which the item occurs in pattern A. According to the logic of statistical pre-emption, the blocking force of went is extremely high because the average speaker has been exposed to thousands of contexts where both went and *goed would have been semantically adequate, but in fact only went was used. From this evidence speakers can quite safely infer that the speech community must have a deep-rooted presumption against *goed, otherwise they would have come across this form given the large number of favorable contexts. The inference in the case of went is supported not only by the high probability of went as opposed to *goed in the input data, but also by the high absolute frequency of went: the higher the absolute frequency of the blocking item, the higher the reliability of the inference. Due to its probabilistic nature, statistical pre-emption can handle in a natural way what we have called in the section headline the “relativity” of the blocking effect. In that respect, alternative mechanisms, such as the Elsewhere Condition, appear as clearly inferior, since they treat blocking as a question of all or none, while in reality all intermediary stages between categorical blocking and free variation can be observed. What a statistical-pre-emption account of blocking also makes clear is that an adequate theory of the phenomenon must be based on a conception of the lexicon as a repository where all words or patterns that we hear, even regular ones, leave memory traces on the basis of which the frequency-based inferences can operate.

3.4. Actual, Possible, and Potential Words

Lexical blocking is generally presented as competition between an actual and a possible word (word form or lexeme), though even this point does not make unanimity.

Followers of Distributed Morphology, for example, do not consider the blocked word to be a possible, well-formed word, since in their model it is not even generated by the grammar, being discarded at lexical insertion by the Elsewhere Principle (cf. Embick & Marantz, 2008). But this account has serious problems anyway. For one, there is the problem of the relativity of blocking mentioned in Section 3.3, “The Relativity of Blocking.” On the other hand, a lexical-insertion account is only able to handle cases of blocking among formations with the same stem. As pointed out by Kiparsky (1983, pp. 13−17) and Rainer (2012), the Elsewhere Principle, which gives precedence to the more specific one of two competing rules, can explain the blocking of -es by -en after ox, or of -er (in the agentive sense) by “zero” after cook, but not that of *goed by went or of *stealer by thief (“blocking by suppletion”), nor that of *gloriosity by glory (“blocking by ancestry”).

However, simply to classify blocked words as “possible” words also falls short of the real complexity of the situation. On the one hand, it is true that thoroughly blocked words sometimes surface under special circumstances, for example, in slips of the tongue, when even competent adult speakers may say things like *goed or *stealer. In child language, as is well known, such forms occur quite commonly as long as the legitimate forms are not yet well entrenched. On the other hand, it has been shown that blocked words behave differently from normal possible words (cf. Giegerich, 2001, p. 77; Rainer, 2012). Normal possible words can be used as bases for further derivation or compounding, just like actual words, while blocked words can’t. The following example may suffice to make the point. A—mostly young—person studying at a school, college, or university is called a student, an irregular derivative which blocks the regular agentive noun *studier that one might expect. This latter derivative is only used in other senses of the verb study, as in Lipsius was a great studier of the stoical philosophy (see The Free Dictionary). Student is, of course, a possible base for further derivations, such as proto student: “The place was swarming with teenagers, most of whom looked like they were clones of the same skinny jeans-wearing, tattooed proto student.” (Kevin Chapman & Richard Bass, Mature: The First Day. Peterborough, U.K.: Kesteven Media, 2014). *Proto studier, by contrast, would be just as odd in this context as *studier itself, although *proto studier cannot be blocked by the occasionalism proto student, which is not part of the general language and hence of the mental lexicon of individual speakers. As we can see, once a word is blocked, further derivation or compounding from this blocked word becomes impossible. It therefore becomes necessary to split the set of possible words into two subsets, ordinary possible words and blocked words, referred to as “potential words” in Rainer (2012). The latter apparently have a distinct status in our mental lexicon, which psycho- and neurolinguists are invited to define more precisely.

3.5. The Locus of Blocking

A last issue that I would like to address is the question of where blocking takes place. It will be convenient to treat lexical blocking and pattern blocking separately.

Pattern blocking is probably best viewed as taking place at the moment of choosing among synonymous patterns for the creation of a new phrase, lexeme, or word form. In this competition, we may surmise, the pattern with the better credentials will win. Among the factors determining the outcome we may mention the number and strength of the features of the base word favoring pattern A over B or vice versa, as well as the memorized formations for each pattern.

The question seems to be more controversial in the case of lexical blocking. As we have already mentioned in Section 3.3, “The Relativity of Blocking,” Hermann Paul conceived of lexical blocking as a struggle between a memorized item and a regular formation. This view of lexical blocking as located in the mental lexicon is still widely held and certainly has some good points on its side, first and foremost, that of being able to account for the correlation between the token frequency of the blocking item and the blocking force: only in a conception of the lexicon as a mental lexicon is the notion of token frequency available at all. It is hardly surprising, therefore, that the most penetrating contributions have come from psycholinguists, especially in the context of the extensive controversy about English past-tense formation (cf. Ambridge & Lieven 2011, Section 5.4; Pinker, 1999, pp. 130−146 and Chapter 6). Theories of grammar, such as Distributed Morphology, where blocking is located at lexical insertion and only takes into consideration the features of the base and the selectional restrictions of the affix, seem to be at a loss to account for the observed frequency effects, quite apart from their incapacity to explain blocking by suppletion and by ancestry, as mentioned in section 3.4, “Actual, Possible, and Potential Words.”

Another question that divides researchers is what it is exactly that blocking prevents. For distributed morphologists, as we have seen in Section 3.4, the blocked word is not even generated by the grammar, and therefore is ungrammatical. The rule itself is blocked. By contrast, in accounts of blocking inspired by Optimality Theory (cf. Kiparsky, 2005), blocking occurs in a grammar-internal selection process after both the blocking item and the blocked one have been generated. In a similar vein, Bauer (1983, p. 88) claimed that “blocking prevents not so much the coining of nonce complex forms as their institutionalization.” He seems to have had in mind an essentially sociological perspective on the phenomenon. We can indeed observe that competing formations sometimes live side by side in a speech community for a long time, even in the idiolect of one and the same speaker, before one wins over the other(s), or semantic or stylistic differentiation takes place. Overabundance seems to be particularly pronounced in relatively heterogeneous speech communities, or in communities with little homogenizing pressure from normative institutions such as academies or schools. The number of derivational doublets, for example, was notoriously higher in Old French or Early Modern English word formation than it is in Modern French or Modern English. This sociological approach to the phenomenon, which takes into consideration the speech community as a whole, is by no means at variance with the psycholinguistic viewpoint mentioned above, which tends to concentrate on the idiolect of single speakers, both views being held together by the Saussurean view of language as a social institution.

4. Critical Analysis of Scholarship

As we have seen, phenomena that have been subsumed here under “blocking” have been the object of linguistic reflection since the very beginning of the study of language. Different authors, however, targeted different subsets of the phenomenon and did so under different labels (to the extent that it received an explicit label at all) and from widely different theoretical backgrounds. As far as the object of study is concerned, scholars concentrated either on inflection, word formation, or syntax, sometimes also on both inflection and word formation, but rarely on the whole range of phenomena. This selective attention has led to partly different terminologies in the respective fields: those specializing in inflection normally prefer the label “suppletion,” which, however, is not a perfect synonym of “blocking,” those concentrating on word formation, “blocking” or “pre-emption,” while in syntax the most common label seems to be “competition.” However, I hasten to add that the correspondence between sub-domain of grammar and label established in the last sentence is not valid for the whole literature. Nevertheless, students who intend to delve more deeply into the problem should be aware of entering a research area covered by moving sands.

The same heterogeneity may also be observed as far as the theoretical background is concerned. One strand of research, going back to Hermann Paul, has concentrated on lexical blocking and tried to devise a theory of the phenomenon on the basis of the mental lexicon, in which concepts such as “memory trace,” “token frequency,” or “productivity” play a crucial role. Interesting recent contributions from this perspective have been provided by psycholinguists (e.g., Maslen et al., 2004) and cognitive linguists such as Goldberg (2011), who uses corpus-linguistic methods to test her theory about how blocking (she speaks of “pre-emption”) can be learned through inferences based on the token frequency of the relevant items. What makes the mental-lexicon approach so attractive is that it can cope quite straightforwardly with what I have called the “relativity of blocking,” that is, its mostly non-categorical nature. Connectionist theories and theories of analogical modeling, which resort to computer simulations, also try to mirror the mental lexicon as closely as possible. By contrast, other accounts of blocking that have been put forward by theoretical linguists operate with more abstract conceptions of the lexicon, as is customary in some quarters of linguistic theorizing. Aronoff (1976), for example, operated with a lexicon organized around stems associated with slots for each canonical meaning that can be regularly derived from that stem. Similar accounts are still in use, at least in studies on inflection—for example in Paradigm Function Morphology. Other conceptions of the lexicon, such as those of Distributed Morphology or Optimality Theory, are still further removed from a mental lexicon. Another dividing line in current accounts of blocking runs between theories that rely on some general principle of synonymy avoidance end those that incorporate the Elsewhere Condition of Pāṇinian descent. It is surprising that this principle still enjoys quite some popularity, although it has been shown repeatedly that it cannot account for several obvious cases of blocking.

As one can see, despite the large body of research that has accumulated, the issue of blocking is still a rewarding avenue of research, located at the center of interesting theoretical questions, but also in need of more large-scale empirical inquiries.

Further Reading

Adams, M. E. (2012). The comparative grammaticality of the English comparative. (Doctoral dissertation). Department of Linguistics, Standford University.Find this resource:

Aronoff, M. (1994). Blocking. In R. E. Asher & J. M. Y. Simpson (Eds.), The encyclopedia of language and linguistics. Vol. 1 (pp. 373–374). Oxford, U.K.: Pergamon.Find this resource:

Carstairs-McCarthy, A. (2010). The evolution of morphology. Oxford, U.K.: Oxford University Press.Find this resource:

Embick, D., & Marantz, A. (2008). Architecture and blocking. Linguistic Inquiry, 39(1), 1–53.Find this resource:

Goldberg, A. E. (2011). Corpus evidence of the viability of statistical pre-emption. Cognitive Linguistics, 20(1), 93–127.Find this resource:

Kiparsky, P. (2005). Blocking and periphrasis in inflectional paradigms. In G. Booij & J. van Marle (Eds.), Yearbook of morphology 2004 (pp. 113–135). Dordrecht, Netherlands: Springer.Find this resource:

Kiparsky, P. (2010). Dvandvas, blocking, and the associative: The bumpy ride from phrase to word. Language, 86(2), 302–331.Find this resource:

Lindsay, M., & Aronoff, M. (2013). Natural selection in self-organizing morphological systems. In N. Hathout, F. Montermini, & J. Tseng (Eds.), Morphology in Toulouse. Selected Proceedings of Décembrettes 7 (Toulouse, December 2–3, 2010) (pp. 133–153). Munich, Germany: LINCOM Europa.Find this resource:

Maslen, R. J. C., Theakston, A. L., Lieven, E. V. M., & Tomasello, M. (2004). A dense corpus study of past tense and plural overregularization in English, Journal of Speech, Language, and Hearing Research, 47(6), 1319–1333.Find this resource:

Rainer, F. (1989). Towards a theory of blocking: The case of Italian and German quality nouns. In G. Booij & J. van Marle (Eds.), Yearbook of morphology 1988 (pp. 155–185). Dordrecht, Netherlands: Foris.Find this resource:

Rainer, F. (2012). Morphological metaphysics: Virtual, potential, and actual words. Word Structure, 5(2), 165–182.Find this resource:

Rainer, F. (2014). Blocking. Oxford Bibliographies in Linguistics.

Schlücker, B., & Plag, I. (2011). Compound or phrase? Analogy in naming. Lingua, 121, 1539–1551.Find this resource:

Veselinova, L. (2014). Suppletion. Oxford Bibliographies in Linguistics.


Adams, M. E. (2012). The comparative grammaticality of the English comparative. (Doctoral dissertation). Department of Linguistics, Standford University.Find this resource:

Ambridge, B., & Lieven, E. V. M. (2011). Child language acquisition: Contrasting theoretical approaches. Cambridge, U.K.: Cambridge University Press.Find this resource:

Aronoff, M. (1976). Word formation in generative grammar. Cambridge, MA: MIT Press.Find this resource:

Aronoff, M. (1994). Blocking. In Asher, R. E., & Simpson, J. M. Y. (Eds.), The encyclopedia of language and linguistics. Vol. 1 (pp. 373–374). Oxford, U.K.: Pergamon.Find this resource:

Bauer, L. (1983). English word-formation. Cambridge, U.K.: Cambridge University Press.Find this resource:

Booij, Geert (2002). The morphology of Dutch. Oxford, U.K.: Oxford University Press.Find this resource:

Brdar, M. (2009). Metonymy-induced polysemy and the role of suffixation in its resolution in some Slavic languages. Annual Review of Cognitive Linguistics, 7, 58–88.Find this resource:

Bréal, M. (1897). Essai de sémantique (science des significations). Paris, France: Hachette.Find this resource:

Carstairs, A. (1988). Some implications of phonologically conditioned suppletion. In G. Booij, & J. van Marle (Eds.), Yearbook of morphology 1988 (pp. 67–94). Dordrecht, Netherlands: Foris.Find this resource:

Carstairs-McCarthy, A. (2010). The evolution of morphology. Oxford, U.K.: Oxford University Press.Find this resource:

Claes, J. (2015). Competing constructions: The pluralisation of presentational haber in Dominican Spanish. Cognitive Linguistics, 26(1), 1–30.Find this resource:

Clark, E. V. (1993). The lexicon in acquisition. Cambridge, U.K.: Cambridge University Press.Find this resource:

Embick, D., & Marantz, A. (2008). Architecture and blocking. Linguistic Inquiry, 39(1), 1–53.Find this resource:

Gardani, F. (2013). Dynamics of morphological productivity: The evolution of noun classes from Latin to Italian. Boston, MA: Brill.Find this resource:

Gardani, F. (2015). Affix pleonasm. In P. O. Müller, I. Ohnheiser, S. Olsen, & F. Rainer (Eds.), Word-formation: An international handbook of the languages of Europe. Vol. 1 (pp. 537–550). Berlin, Germany: de Gruyter Mouton.Find this resource:

Gauger, H. M. (1972). Zum Problem der Synonyme. Tübingen, Germany: Narr.Find this resource:

Giegerich, H. J. (2001). Synonymy blocking and the elsewhere condition: Lexical morphology and the speaker. Transactions of the Philological Society, 99(1), 65–98.Find this resource:

Goldberg, A. E. (2011). Corpus evidence of the viability of statistical pre-emption. Cognitive Linguistics, 20(1), 93–127.Find this resource:

González-Díaz, V. (2008). English adjective comparison. A historical perspective. Philadelphia, PA: John Benjamins.Find this resource:

Janda, R. D., & Sandoval, M. (1984). “Elsewhere” in morphology. Bloomington, IN: Indiana University Linguistics Club.Find this resource:

Kiparsky, P. (1983). Word-formation and the lexicon. In F. Ingemann (Ed.), Proceedings of the 1982 Mid-American Linguistics Conference (pp. 3–22). Lawrence: University of Kansas.Find this resource:

Kiparsky, P. (2005). Blocking and periphrasis in inflectional paradigms. In G. Booij, & J. van Marle (Eds.), Yearbook of morphology 2004 (pp. 113–135). Dordrecht, Netherlands: Springer.Find this resource:

Kiparsky, P. (2010). Dvandvas, blocking, and the associative: The bumpy ride from phrase to word. Language, 86(2), 302–331.Find this resource:

Levin, S. (1972). Non-paradigmatic forms: Suppletion or pre-emption. Foundations of Language, 8(3), 767–811.Find this resource:

Lindsay, M., & Aronoff, M. (2013). Natural selection in self-organizing morphological systems. In N. Hathout, F. Montermini, & J. Tseng (Eds.), Morphology in Toulouse. Selected Proceedings of Décembrettes 7 (Toulouse, December 2–3, 2010) (pp. 133–153). Munich, Germany: LINCOM Europa.Find this resource:

McCawley, James D. (1978). Conversational implicature and the lexicon. In P. Cole (Ed.), Syntax and Semantics 9: Pragmatics (pp. 245–259). New York: Academic Press.Find this resource:

Malicka-Kleparska, A. (1985). Parallel derivation and lexicalist morphology: The case of Polish diminutivization. In E. Gussmann (Ed.), Phono-morphology: Studies in the interaction of phonology and morphology (pp. 95–112). Lublin, Poland: Redakcja Wydawnictw Katolickiego Uniwersitetu Lubelskiego.Find this resource:

Marle, Jaap van (1986). The domain hypothesis: The study of rival morphological processes. Linguistics, 24, 601–627.Find this resource:

Maslen, R. J. C., Theakston, A. L., Lieven, E. V. M., & Tomasello, M. (2004). A dense corpus study of past tense and plural overregularization in English. Journal of Speech, Language, and Hearing Research, 47(6): 1319–1333.Find this resource:

Miyagawa, S. (1984). Blocking and Japanese causatives. Lingua, 64(2–3), 177–207.Find this resource:

Mondorf, B. (2009). More support for more-support. The role of processing constraints on the choice between synthetic and analytic comparative forms. Philadelphia, PA: Benjamins.Find this resource:

Osthoff, H. (1899). Vom Suppletivwesen der indogermanischen Sprachen. Heidelberg, Germany: Hörning.Find this resource:

Paul, H. (1920). Prinzipien der Sprachgeschichte. 5th ed. Halle, Germany: Niemeyer.Find this resource:

Paul, H. (1896). Ueber die Aufgaben der Wortbildungslehre (pp. 692–713). Munich: Germany: F. Straub.Find this resource:

Pinker, S. (1999). Words and rules: The ingredients of language. London, U.K.: Weidenfeld & Nicolson.Find this resource:

Plag, I. (1999). Morphological productivity: Structural constraints in English derivation. Berlin, Germany: Mouton de Gruyter.Find this resource:

Plank, F. (1981). Morphologische (Ir-)Regularitäten. Tübingen, Germany: Narr.Find this resource:

Rainer, F. (1989). Towards a theory of blocking: The case of Italian and German quality nouns. In G. Booij & J. van Marle (Eds.), Yearbook of morphology 1988 (pp. 155–185). Dordrecht, Netherlands: Foris.Find this resource:

Rainer, F. (2012). Morphological metaphysics: Virtual, potential, and actual words. Word Structure, 5(2), 165–182.Find this resource:

Schindler, J. (1972). Wortbildungsregeln. Wiener Linguistische Gazette, 1, 39–52.Find this resource:

Schlücker, B., & Plag, I. (2011). Compound or phrase? Analogy in naming. Lingua, 121, 1539–1551.Find this resource:

Thornton, A. M. (2012). Reduction and maintenance of overabundance: A case study on Italian verb paradigms. Word Structure, 5, 183–207.Find this resource: