How adjectival can a participle be?

Subsective Gradience in English 2nd Participles

by Samirah Aljohani (Author)
©2022 Monographs XVI, 422 Pages


«This book takes theoretical linguistics by storm, moving our understanding of the passive construction onto a whole new level. Samirah Aljohani puts the adjectival passive under the empirical lexico-grammatical microscope, producing numbers which both dazzle and clarify. Inspired science from copious data presented in an accessible style – absolutely brilliant!» (Dr Christopher Beedham, University of St Andrews, Scotland)
Most analyses of the English passive (formed with be + V-ed) claim that there is a verbal passive and an adjectival passive. How can the same form express polar opposite meanings? This study of the adjectival passive reconciles the contradiction using Christopher Beedham’s aspect analysis of the passive, in which the so-called actional passive (verbal passive) is said to express an action and its resultant state.
In the study, the author presented approximately one thousand 2nd participles, mainly from transitive verbs, to three native speaker informants in putative noun phrases such as an accepted practice and putative clauses with un-, such as It is unaccepted, and asked the informants to say if they are grammatical, ungrammatical or borderline. She also interrogated her participles in the British National Corpus for their adjectival properties. In this way, she arrived at five adjective-like properties which a 2nd participle can have. Finally, she put her participles into eight groups, ranging from «0% state, 100% action» to «50% state, 50% action», depending on how many and which of the five adjective-like properties they can exhibit. The result is a new gradient scale of adjectival passives.

Table Of Contents

  • Cover
  • Title
  • Copyright
  • About the author
  • About the book
  • This eBook can be cited
  • Contents
  • List of Figures
  • List of Tables
  • Preface
  • Abbreviations and Symbols
  • Chapter 1 Introduction
  • 1.1 The Adjectival Passive
  • 1.2 Research Questions
  • 1.3 Objectives of the Study
  • 1.4 What Is the Adjectival Passive?
  • 1.5 Implicit Arguments in Adjectival Passives, and Types of Adjectival Participles
  • 1.6 Summary
  • Chapter 2 The Aspect Analysis of the Passive
  • 2.1 Passive Voice and Transitivity
  • 2.2 Aspect
  • 2.3 The Aspect Analysis of the Passive
  • 2.4 Lexical Aspect and Telicity
  • 2.5 Obligatory Modification in Passive and 
Attributive Participles
  • 2.6 The Need for a Gradient Scale
  • 2.7 Theoretical Approach and Rationale
  • 2.8 Summary
  • Chapter 3 Methodology: Structuralism, Exceptions and Corpora
  • 3.1 Saussure’s Structuralism
  • 3.2 Langue and Parole
  • 3.3 Methodology
  • 3.4 The Research Design
  • 3.5 Methodological Considerations
  • 3.6 Conclusion
  • Chapter 4 A Correlation between Attributive 2nd Participles and Adjectival Passive
  • 4.1 Adjectival Passive, Attributive Participles and the Aspect Analysis
  • 4.2 The First Investigation
  • 4.4 Conclusion
  • Chapter 5 Subsective Gradience in 2nd Participles
  • 5.1 Participles in Adjectival Passives
  • 5.2 Form and Function Dichotomy
  • 5.3 Grammatical Gradience
  • 5.4 Groups of 2nd Participles
  • 5.5 2nd Participles as a Case of Subsective Gradience as Opposed to Intersective Gradience
  • 5.6 Concluding Remarks
  • Chapter 6 Recap and Implications
  • 6.1 Un-​ prefixation
  • 6.2 The Interpretation of Attributive 2nd Participles
  • 6.3 Transitivity and Passivisation
  • 6.4 The Adjectival Passive in the Aspect Analysis
  • 6.5 Conclusion
  • Chapter 7 Summary and Conclusion
  • 7.1 Summary of the Book
  • 7.2 Conclusion
  • Bibliography
  • Appendices
  • Appendix 1 The Sample of 2nd Participles Investigated
  • Appendix 2 Questionnaire A
  • Appendix 3 Questionnaire B
  • Appendix 4 Group A
  • Appendix 5 Group B
  • Appendix 6 Group C
  • Appendix 7 Group D
  • Appendix 8 Group E
  • Appendix 9 Group F
  • Appendix 10 Group G
  • Appendix 11 Group H
  • Appendix 12 Exceptions to Group H
  • Appendix 13 Examples Extracted from the BNC
  • Appendix 14 Examples Extracted from the BNC and COCA
  • Index
  • Series Index

←x | xi→


Table 2.1. Frequencies of the passive progressive (and the perfect progressive) in the BNC

Table 3.1. The morpho-​syntactic variables investigated in the BNC (Appendices 4–​12)

Table 3.2. Examples from the BNC (Appendix 13)

Table 4.1. Inconclusive responses (Questionnaire A)

Table 4.2. Questionnaire A: Results of attributive function of participles

Table 4.3. Questionnaire B: Results for adjectival passive

Table 4.4. Cross-​tabulation of attributive function and adjectival passive in 2nd participles

Table 4.5. Kappa statistics

Table 4.6. Ungrammatical attributive participles that form grammatical adjectival passives

Table 4.7. Participles that are grammatical in adjectival passive but borderline attributively

Table 4.8. Part of Table 4.4 reproduced after examining groups 1 and 2

Table 5.1. Groups of 2nd participles based on morphological and syntactic properties

Table 5.2. Participial groups based on subsective gradience

Table 5.3. Zero state participles

Table 5.4. Group B participles

Table 5.5. Prepositional passive participles

←xi | xii→

Table 5.6. Attributive-​only 2nd participles

Table 5.7. Verbal passive participles

Table 5.8. Un-​ prefixation and adverb modification of attributive participles in group E

Table 5.9. Group F participles

Table 5.10. Group G participles

Table 5.11. Un-​ prefixation and adverb modification attributively of participles in group G

Table 5.12. Group I participles

Table 5.13. Un-​ prefixation and adverb modification of attributive participles in group H

Table 5.14. Exceptions to group H

←xii | xiii→


This book was originally a doctoral thesis. I would like to express my sincere gratitude to all the people who supported me. Thanks are due first to my supervisor, Dr Christopher Beedham. Words cannot thank you enough. The academic expertise you provided and the moral support was more than I hoped for. Your constant encouragement and enthusiasm were priceless. The freedom to conduct research and voice my thoughts enabled me to grow as a researcher. I could not hope for a better adviser and mentor for my PhD.

Thanks and gratitude go to my examining committee, Dr Kormi Anipa (University of St Andrews) and Dr Andrew McIntyre (Humboldt-​Universität zu Berlin), for their feedback and assistance. Thank you for your key insights, input and vital critiques, which have significantly made this work a better thesis. Thanks are also due to Taibah University, which funded me throughout my PhD studies.

Thanks extend also to all my colleagues and friends, past and present, for constant encouragement and support. A big, heartfelt thank you goes to all my family, who always encouraged and supported me in all my pursuits. Thank you so much for believing in me and helping me in whatever way they could during this challenging time of my life.

←xvi | 1→

Chapter 1


1.1 The Adjectival Passive

This book discusses the adjectival passive in English using an aspect analysis of the passive (Beedham 2014, 2005, 1982). There are two widely recognised types of passive: the so-​called verbal passive, formed with an auxiliary verb and a 2nd participle,1 be +​ V-​ed, as in the door was locked by the janitor, and the adjectival passive, also formed with a 2nd participle, as in the door seemed locked, which is considered to express a state resulting from a prior event. The adjectival passive is so-​called because it shows behaviour typical of adjectives, and has a more stative meaning, similar to adjectives: the participle can be modified with very and too, and accepts prefixation with un-​; the verb to be can be replaced by seem, look, appear and remain. Another feature of the adjectival passive is that the by-​phrase is often inadmissible. In this book I do not consider that the attributive use of the passive participles, as in, for example, a damaged car, is a defining property of adjectival passives, as some linguists do. Attributive use makes the participle adjective-​like, but it does not make it an ‘adjectival passive’ in the strict sense of the term as defined in this book. The reason has to do in part with the main topic of the book, which is that in the 2nd participle there is not a black and white dichotomy between action (or more accurately ‘action +​ state’ in prototypical passives) in so-​called verbal passives and state in adjectival passives, but rather a continuum of ←1 | 2→different mixes and proportions of action and of state, presented empirically and discussed in detail in Chapter 5.

There are three key issues related to the adjectival passive: its formation, whether it has evidence of voice, and its relation to attributive participles. Some of the literature on the adjectival passive debated whether it is formed in the syntax or the lexicon, for example, Emonds (2006), Embick (2004), Wasow (1977), Levin and Rappaport (1986). In some studies, it is viewed as a copular plus predicative complement, for example, in Wanner (2009: 13) and Huddleston and Pullum (2002: 1,436).

Voice in the sense of Kratzer (1996) is a VoiceP, a high verbal projection head which introduces the external argument and is ‘severed’ from the verb and realised outside the VP. There is debate over whether adjectival participles in any sense refer to an agent or other external arguments in their semantic representations. A related question is whether adjectival participle formation explicitly involves the suppression of such arguments. This was usually denied in earlier studies (Embick 2004; Kratzer 2000, 1996, 1994), but some recent studies, mainly in the generative paradigm, have called this into question (Anagnostopoulou 2017, Alexiadou et al. 2015, Bruening 2014, and McIntyre 2013). These studies express this in terms of a claim that the participles involve some variant of Kratzer’s Voice head, but one could imagine the same idea expressed in non-​syntactic terms in other theories. For example, Bruening (2014) suggested an AdjP head for English resultant participles which stativises the verbal projection; Alexiadou et al. (2015) proposes that the AspP, embedded under an adjectival little a for Greek, arguing that the presence vs absence of AspP explains the presence (in English and German) or absence (in Greek) of restrictions on modifiers.

The attributive function of 2nd participles is important because their prenominal position is associated more with adjectives: the damaged car (2nd participle), the red car (adjective). The attributive function is controversial in that we must ask, what is its relation to adjectival passives?2 ←2 | 3→Do we consider attributive 2nd participles as instances of the adjectival passive, because they appear in a position typically occupied by adjectives, as Bruening (2014), Embick (2004), Bresnan (1995), Levin and Rappaport (1986) and Wasow (1977) do? If they are not adjectival passives, we would have to make a difficult decision: do we consider them to be reduced relatives, in the way that McIntyre (2013), Sleeman (2007), Laskova (2007) and Laczkó (2001) consider some of the attributive participles to be eventive participles, or do we argue, in line with Huddleston and Pullum (2002) and Quirk et al. (1985), that they are participial adjectives, that is, first and foremost adjectives?

2nd participles are typically viewed as derived verbal forms that function like adjectives (Crystal 2008; Huddleston and Pullum 2002; Quirk et al. 1985), which is precisely why constructions involving them defy analysis: are they verbs or adjectives (or both)? To illustrate the divided position on 2nd participles, the Oxford English Dictionary (OED) lists allowed, damaged and broken in separate entries as adjectives, while enjoyed, consisted, forgotten and replied are not listed as adjectives, yet complained is listed as an adjective, but not as a separate entry, but under the verb to complain.3 What motivates such decisions? Though ours is not a lexicographical study the examples from the OED show the complex nature of 2nd participles.

What we are in dire need of is an empirical survey of 2nd participles which identifies which participles can form an adjectival passive (in the sense of this book), which can function attributively, which can do both, and which cannot do both. When that is accomplished it will help us understand the nature of the relation between the adjectival passive as a clause and the attributive function of participles. Our study attempts to investigate this relationship empirically, and to explain the adjectival passive within an aspectual analysis. One of the major difficulties that non-​native speakers of English (like myself) face is that the ability of participles to function attributively and to form adjectival passives is not something that is made explicit and recorded in dictionaries. The literature on the adjectival ←3 | 4→passive does not provide us with such detailed information. I hope that by surveying 1,043 2nd participles in the way described in this book we will fill a gap in the study of the adjectival passive empirically.

Finally in this section, I will clarify the main terms that are used here. ‘Attributive participles’ are participles which modify a noun in prenominal position, for example, a damaged car.4 The term ‘verbal passive’ is used here to refer to a clausal passive of the form be +​ V-​ed which either has a by-​phrase or can take a by-​phrase, as in he was murdered (by the butler). The term ‘adjectival passive’ is used here for a passive clause in which the participle has any of the following three formal properties normally seen with adjectives: (i) the verb to be is replaced by remain, look, seem, etc., for example, the car looked damaged; (ii) the participle is prefixed with un-​, for example, he was unhurt; (iii) the participle is graded with very or too, for example, she was very amused. There are cases where the passive is ambiguous, for example, the car is damaged, and only a wider context specifies whether it is adjectival or verbal. The adjectival passive of the un-​ type, for example, he was unhurt, is examined in Chapter 4 using grammaticality judgements from native speaker informants, whilst the corpus investigation in Chapter 5 looks in detail at adjectival passives with the three above-​mentioned formal properties. ‘Adjective’ is a form label, but we use ‘adjectival’ or ‘adjective-​like’ to refer to the grammatical function. ‘State’ is a semantic term frequently associated with the meaning of adjectives, in contrast with event or action. ‘State’ is also used in the book to refer to the lexical aspect of ‘stative verbs’ (e.g. know, love, like which are incompatible with the progressive aspect), or the aspect of the clause as a whole, for example, Mary was sick. ‘State’ is discussed in Chapter 2 in aspect theory, but is discussed again in Chapter 5 along with adjective in terms of gradience theory.←4 | 5→

1.2 Research Questions

Within an aspectual theory, this study attempts to answer the following questions:

1.What is the relationship between the adjectival passive and the attributive function of 2nd participles?

2.Is there a gradient scale by which 2nd participles show different mixes of action and state that can be grammatically manifested as varying degrees of adjective-​like behaviour in 2nd participles?

1.3 Objectives of the Study

The key issue in investigating the adjectival passive is, first, to understand statistically what the relation between adjectival passive clauses and the attributive function is, and, second, to find out in data-​driven detail what the adjective-​like behaviour of participles means, and how actually ‘adjectival’ participles are. We use two methods of data collection –​ obtaining grammaticality judgements from native speaker informants, and extracting examples from a large corpus, in our case the British National Corpus (BNC) –​ which are the empirical underpinning of our ideas, the quantitative foundation of a qualitative analysis. We will call the two methods ‘investigation’, that is, collecting data and analysing it, ‘the first investigation’ and ‘the second investigation’, respectively. The two investigations are two empirical strands which run through the whole of this book.

This study seeks to investigate the adjectival passive within an aspectual analysis of the passive. One of the major issues identified in the studies of the adjectival passive reviewed in Sections 1.4 and 1.5 is that no wide-​scale and detailed empirical study of the adjectival passive seems ever to have been carried out. The ability of a participle to form an adjectival passive is not a feature that is noted in dictionaries. Furthermore, the exact nature of ←5 | 6→the relationship between attributive function and adjectival passive is controversial. Attributive participles are either argued to be adjectival passives, or said not to be adjectival passives. Where attributive participles are not considered adjectival passives, they are analysed either as participial adjectives –​ with adjectival status –​ or as reduced relatives –​ with verbal status.

This study aims to investigate empirically the adjectival passive in relation to the attributive function of the relevant participles. If the verbal passive is argued to show an aspectual behaviour, and the attributive function of participles can be due to telicity or informativeness considerations, as discussed in Sections 2.4 and 2.5, then the lexical aspect of the verb from which participles are derived would have an effect on the ability of participles to form the adjectival passive. If such an association between attributive participles, adjectival passive and verbal passive exists, then it strongly suggests that an aspectual approach, in which telicity is an optimal condition, is relevant to the adjectival passive.

The study takes the view that the lexical aspect of the verbs from which participles are derived has an effect on the adjectival and the adjective-​like function of participles. The controversial status of the adjectival passive and attributive participles is due to the fact that participles mirror the syntactic distribution of adjectives. In Chapter 4 the association between attributive participles and the adjectival passive shows that most participles in the adjectival passive can function attributively, though some require obligatory modification. The second investigation, presented in Chapter 5, is a corpus-​based study which uses the BNC to try to capture the adjective-​like function of participles on a gradient scale, a continuum, indicative of or imply a meaning ‘action +​ state’ as a linguistic sign. The second investigation surveys 1,035 2nd participles across three constructions: adjectival passive, verbal passive and attributive function in noun phrases (NPs).5 I will focus on form and function as they are empirically inferred from the data of the corpus.←6 | 7→

1.4 What Is the Adjectival Passive?

The adjectival passive is a construction in which the 2nd participle exhibits an overlap between adjectives and verbs. It is referred to as a passive clause with an adjectival, statal meaning. Huddleston and Pullum (2002: 1,436) restrict the term ‘adjectival passive’ to a 2nd participle that is used as a predicative complement, as in they are married. They consider the entire clause to be a complex-​intransitive clause that contains but is not itself an adjectival passive. In generative accounts, the term ‘adjectival passive’ tends to refer to the 2nd participle that is used both predicatively and attributively, as Levin and Rappaport (1986: 626) affirm: ‘any passive participle found in prenominal position is therefore also adjectival, not verbal’. As Levin and Rappaport acknowledge, unfortunately, this approach excludes attributive participles with perfect interpretation, for example, an escaped criminal.

What sets the adjectival passive apart from the verbal passive is the existence of certain syntactic and morphological indicators, adumbrated in Section 1.1 above. These indicators, or tests, have often been cited in previous studies: McIntyre (2013), Wanner (2009), Toyota (2009), Emonds (2006), Embick (2004), Huddleston and Pullum (2002), Quirk et al. (1985), Levin and Rappaport (1986), Wasow (1977), Hust (1976, 1978) and Siegel (1973). In the earlier accounts, Hust (1976) and Siegel (1973) placed special emphasis on un-​ prefixation. They called the construction the ‘unpassive’. To clarify the syntactic tests for adjectival passive presented in Section 1.1 I will briefly review here the tests as given in Huddleston and Pullum (2002: 1,436). They mention modification with very and too, the possibility of replacing the verb to be with other verbs, for example, seem, look, remain and predicative un-​ participles.

In this study I restrict the term ‘adjectival passive’ to a passive clause that has any of these three features, discussed further in Sections 1.4.1, ←7 | 8→1.4.2 and 1.4.3. As alluded to at the beginning of Section 1.1, I exclude 2nd participles used attributively from being considered adjectival passives, for reasons that I will elaborate on in Section 1.4.4.

1.4.1 Modification with very and too

One of the tests provided by Huddleston and Pullum (2002) for the adjectival passive is modification with very and too, for example, they were very worried, he was too embarrassed by their behaviour to acknowledge that he was their son (Huddleston and Pullum 2002: 1,436). Certain ungradable verbs, for example, calculate, estimate, rate, represent, behave and manage, can be made gradable when prefixed with over-​, under-​ or mis-​, as in miscalculate, over/​underestimate, over/​underrate, misrepresent, misbehave and mismanage (Quirk et al. 1985: 595). The following are examples from the BNC:

1.I think that bay is very underrated.

2.Most housewives are very undervalued.

3.Social services were too overworked to be accessible.

4.… but we’re very misrepresented.

Not all adjectives are gradable, and neither are all participles. The presence of very suggests a gradable quality, but very does not occur with all 2nd participles.6 For example, very is incompatible with married, although married can still form an adjectival passive, as in she was unmarried. Modification with very and too indicates that the passive participle is adjectival, thus we expect it to be incompatible with voice-​related modifiers. But that turns out not to be the case –​ see below.

←8 | 9→Certainly very and too in a passive clause would encourage an adjectival interpretation, and yet there is evidence that by-​phrases are compatible with very/​too. The presence of a by-​phrase suggests a verbal reading, but the degree modifier suggests a state or property or adjectival reading. The following example, cited in Huddleston and Pullum (2002: 1,436), illustrates this:

5.He was too embarrassed by their behaviour to acknowledge that he was their son.

Huddleston and Pullum say that embarrassed is an adjectival passive, because the participle is modified by too. The issue that arises from having both too and a by-​phrase is that too requires an adjective-​like state or a quality to grade and modify, yet the by-​phrase requires that the NP is an agent of an action. There are many examples of the adjectival passive in which a by-​phrase is admissible.

1.4.2 Auxiliary be

One test for judging whether a passive is verbal or adjectival is the ability to replace be with verbs like seem, look, smell, sound, become, act and remain. These verbs usually select adjectives as complements, for example, she looks beautiful. In the adjectival passive the complement is a 2nd participle. This is a popular test for adjectival passives that is cited in Wanner (2009), Toyota (2009), Huddleston and Pullum (2002), Levin and Rappaport (1986), Quirk et al. (1985) and Wasow (1977), to name just a few. As just shown to be the case with very and too, by-​phrases are also admissible here sometimes, as in the cat remained unnoticed by the guests for several minutes (Huddleston and Pullum 2002: 1,437). Although remained encourages a statal reading of the passive, the sentence has a by-​phrase.

1.4.3 Un-​ prefixation

The third, and most important, test for the adjectival passive is un-​ prefixation. Un-​ prefixes with adjectives to give a negative meaning, for example, ←9 | 10→unclear (not clear), unhelpful (not helpful), whilst the un-​ prefix with verbs provides a reversal meaning, for example, to unbutton, to unlock.7 2nd participles as non-​finite verb forms mostly permit negative un-​, for example, unbroken, but a few allow reversal un-​, for example, uncovered, in which case ambiguities can arise, as in the suitcase was unpacked. The ambiguity is due to the structure of the prefixed participle: negative un-​ attaches to the participle [A un [A [V pack] ed]], meaning not in a packed state resulting from a prior event, whereas reversal un-​ attaches to the verb base [A [V unpack] ed], giving an eventive reading in which the reversal of the event of packing was carried out. The importance of un-​ prefixation is that it is a test of the adjectival status of a given word, and in terms of 2nd participles, Levin and Rappaport (1986: 626) argue that ‘passive participles that are prefixed with un-​ are categorially adjectival and never verbal’. Un-​ prefixation with 2nd participles is fairly productive, as it is with adjectives (Huddleston and Pullum 2002: 1,440; Quirk et al. 1985: 1,540).

6.The cat was unnoticed by the guests for several minutes. (Huddleston and Pullum 2002: 1,440)

7.We were unimpressed by his attempts. (Quirk et al. 1985)

It could be argued that example 6 is an adjectival construction, and the semantics of the participles unnoticed and unimpressed is ‘not noticed’ and ‘not impressed’ (comparable to unclear). Prefixation with un-​ suggests that there is a statal meaning in the 2nd participle, which is a striking ability of passive participles. Not only that, although the participles above are supposed to be statal in meaning, we also have agentive by-​phrases.

Wasow (1977: 341), citing some of Siegel’s (1973) examples, argues that there is a clear contrast between adjectival and verbal passive participles. He notes that some 2nd participles prefix with un-​ but only when they are passive, not active (perfect). It is true that un-​ prefixation is generally more productive with passive 2nd participles. Un-​ prefixation is permitted with 2nd participles derived from intransitive verbs, for example, unfallen, unmelted and unexploded, which can be based on unaccusative verbs, for example, ←10 | 11→unfallen leaves, unarrived guests (McIntyre 2013: 34), though they are not as productive as passive participles. Attributively un-​ prefixation is allowed in both a passive meaning, as in the unnoticed cat, and a perfect meaning, as in unfallen leaves. Also, predicatively un-​ attaches to participles derived from both intransitive verbs, as in the bed looks unslept in, and transitive verbs, as in the cat was unnoticed.

To come back to Wasow (1977: 339), let us examine the following examples. Wasow, discussing Siegel’s (1973) paper, argued that when grammatically correct un-​passives are converted into active clauses, prefixation with un-​ is ungrammatical. Some of Wasow’s (1977: 339) examples are given below:

8.Our products are untouched by human hands. *Human hands untouch our products.

9.The island was uninhabited by humans. *Humans uninhabited the island.

He argues that prefixation with un-​ is only possible when sentences are in the passive voice. He argues (1977: 340) that the peculiarity of the admissibility of un-​ with the passive participles seen above can be removed if we consider the participles lexical adjectives, because it is not unusual to have an adjective prefixed with un-​, for example, unhappy. He argued that these are sentences formed from the lexicon and not by means of transformational rules. There are no verbs such as *untouch or *uninhabit. His analysis has been cited in many studies, as it provides a solution for the peculiar adjectival passives, although it fails to account for the passivisation of 2nd participles like slept in and the non-​passivisation of transitive verbs such as resemble and equal.

Another piece of evidence for un-​ generally favouring passive participles is what Embick (2004) calls ‘voice reversals’, in which 2nd participles have a perfect interpretation when used in attributive position, as in a confessed criminal, that is, a criminal who has confessed. Un-​ prefixation on 2nd participles in the attributive is usually only allowed if the resulting NP interpretation is passive, for example, *an unconfessed killer is ungrammatical but an unconfessed crime is grammatical. Such participles do not have this restriction if used without un-​ prefixation and modify a noun that is an agent, as in a confessed killer (see Section 1.4.4 for a claim in Bruening (2014) that killer is an internal argument).

←11 | 12→I believe that prefixation with un-​ is the most important test and a clear-​cut one of the adjectival status of a passive participle. Grading with degree modifiers such as very and too is more restricted, as discussed above. Whether the auxiliary be can be replaced with other copular verbs depends on the 2nd participle under investigation and the lexical aspect of the copular verbs. I believe that of the three features, modification of participles with very or too, replacing be with remain, seem, look and un-​ prefixation, un-​ prefixation is the most productive test, but the matter is nevertheless complex. Un-​ prefixation enables a passive to be interpreted adjectivally and enables certain participles to occur in attributive position, as with read and seen, and it has implications for voice, as affirmed by Embick (2004). The study by Alexiadou et al. (2015: 169) indicates that manner-​/​agent-​oriented adverbials are not compatible with an adjectival passive when the 2nd participle is prefixed with un-​, although there is evidence of voice elsewhere in the adjectival passive.8 The reason for this, in agreement with Embick (2004), is related to scope and is not for lack of voice, because examples that have both un-​ prefixation and by-​phrases are not infrequent.9

In the first investigation (Chapter 4) we investigate adjectival passive with un-​, whilst in the second investigation (Chapter 5) we look for adjectival passives with any of the three features discussed in Sections 1.4.1, 1.4.2 and 1.4.3, including un-​.←12 | 13→

1.4.4 Occurrence in prenominal position: Attributive function

The status of 2nd participles in attributive function10 has led to divided opinions. Bruening (2014), Embick (2004), Levin and Rappaport (1986) and Wasow (1977) call a participle an adjectival passive if it occurs as ‘a prenominal modifier’. However, other recent studies have taken a different approach, in which attributive participles are not adjectival passives. McIntyre (2013: 22–​23) excludes the internal use of participles in a determiner phrase (DP) as a distinguishing test for the adjectival status of participles.11 Embick (2004: 357) comments on participles used attributively, arguing that their meaning cannot be tested because of the structure of the DP which bars prenominal modifiers with post-​head material, and further, that 2nd participles are assumed in the literature to be adjectival when used attributively for ‘distributional reasoning: only adjectives occur prenominally’. Laczkó (2001: 7) considers attributive participles not to be adjectival, arguing when discussing a fallen leaf and an opened can that both fallen and opened ‘are as dynamic as their input verbs’. On the other hand, Huddleston and Pullum (2002) and Quirk et al. (1985) also do not consider attributive participles to be adjectival passives, but consider them participial adjectives. The car is undamaged is an adjectival passive, while a damaged car and an undamaged car are NPs with 2nd participles functioning attributively, as participial adjectives.12

←13 | 14→In the present book, I consider the attributive 2nd participle to be related to the adjectival passive in that they both indicate some state meaning, but I do not include attributive 2nd participles under our concept of ‘adjectival passive’. Following McIntyre (2013), Emonds (2013) and Huddleston and Pullum (2002) among others, in this study the adjectival passive is considered to be a passive clause in which either un-​ is prefixed to the 2nd participle, or it has very or too modifying the 2nd participle, or it has be replaced by an auxiliary verb like remain or seem, as discussed in Sections 1.4.1, 1.4.2 and 1.4.3.

In studies such as Bruening (2014), Embick (2004), Levin and Rappaport (1986) and Wasow (1977) the use of a 2nd participle in attributive function is used as syntactic evidence of the adjectival status of the participle and, occasionally, as an actual example of an adjectival passive. However, putting the adjectival passive and the 2nd participles used in attributive function in the same category creates confusion and problems.13 Below are some of Bruening’s (2014: 380–​381) examples, in respect of which he argued that adverbs of manner appear in adjectival passives:

10. a.… for he seems dressed with more studied elegance than anybody here.

b.The band in the Metropole finished up a clumsily danced cha cha and a singer came on. (Started Early, Took My Dog, Kate Atkinson)

c.A poorly played hand in No-​Limit can cost you your entire stack, while a well-​played hand can double up your stack.

d.… to be conscious of complete truthfulness while telling carefully constructed lies, … .

Bruening argues that here there is evidence for manner/​agent-​oriented adverbials that modify adjectival passives. 10a does not have an ‘adverb of manner’, but the underlined PP is a ‘manner modifier’ (PPs tend to be more compatible with adjectives). While examples 10b–​10d have adverbs of manner, the adjectival participles are attributive. I am not alone in ←14 | 15→noticing that the examples he cited are all in attributive function, except for 10a. Alexiadou et al. (2015: 172) argue:

The question is whether examples a-​d indeed show what they are supposed to show given that attributive participles have been argued in the literature to be ambiguous between true adjectival participles and reduced relatives based on passives.

Another reason for restricting the term ‘adjectival passive’ to the morpho-​syntactic structure that has one of our above-​named features is the ability of certain intransitive 2nd participles to occur attributively, as in an escaped prisoner, which has a perfect interpretation and not a passive interpretation. The interpretation of participles when used attributively is considered to be related to whether the participle is used in the active or passive voice (Huddleston and Pullum 2002: 541–​542; Quirk et al. 1985: 1,327–​1,329), with the passive interpretation being more frequent and expected than the perfect interpretation.14 Furthermore, some 2nd participles can be either with a perfect or a passive interpretation in NPs. Without un-​ prefixation the attributive participle has either passive or perfect interpretation, but when modified with un-​, as in confessed, it has to be passive: an unconfessed crime, but not *an unconfessed criminal. Interestingly, Embick (2004: 360) categorises such participles as a case of voice reversal. Attributively, it is grammatical to say a confessed criminal (perfect interpretation), and a confessed crime (passive interpretation), but *an unconfessed criminal (perfect interpretation) is ungrammatical. Nevertheless, reversing voice from active to passive as in an unconfessed crime makes un-​ prefixation attributively grammatical. Bruening (2014: 416–​417) discusses similar participles to confessed, for example, recanted, declared, committed, but differs from Embick in that he says un-​ prefixation is allowed in an unconfessed killer, an unrecanted Chomskyan. Bruening argues that killer is not an ←15 | 16→external argument, that is, the reading ‘a killer who has confessed’ is not right, but actually means ‘someone who has confessed to being a killer … an internal argument of the participle’.15

As regards lexical category, attributive function is a canonical function of adjective, thus 2nd participles in this function are problematic. Huddleston and Pullum (2002: 541–​542) –​ though they generally consider attributive participles to be participial adjectives –​ say that if an attributive participle cannot function as a predicative adjective, then the participle is a verb, as in their examples:

11.A rarely heard work by Purcell  (cf. *the work by Purcell is rarely heard)

12.The murdered man       (cf. *the man is murdered)

13.A failed businessman      (cf. *the businessman is failed)

14.The recently departed guests   (cf. *the guests are recently departed)


XVI, 422
ISBN (Softcover)
Publication date
2022 (May)
aspect analysis adjectival passive verbal passive attributive participle obligatory modification requirement informativeness passivisation telic How adjectival can a participle be? Samirah Aljohani
Oxford, Bern, Berlin, Bruxelles, New York, Wien, 2022. XVI, 422 pp., 5 fig. col., 4 fig. b/w, 25 tables.

Biographical notes

Samirah Aljohani (Author)

SAMIRAH ALJOHANI is Assistant Professor of Linguistics at Taibah University, Medina, Saudi Arabia. She completed her Masters at University College London (UCL) and her PhD at the University of St Andrews, Scotland.


Title: How adjectival can a participle be?