A corpus-based analysis of light verb constructions with MAKE and DO in British English

. The comparison of light verb constructions with the verbs make and do has not received much attention. The present paper is an attempt to contribute to the study of these constructions. It aims to analyze the light verbs make and do in combination with the same deverbal nouns to identify similarities and differences between the two light verbs and examine the contribution of the light verbs and deverbal nouns to the light verb construction in terms of semantic and syntactic features. The research is corpus-based, and the data for analysis are collected from the British National Corpus (BNC). The semantic description deals with characteristics such as generality, polysemy and aspectual meaning, and the syntactic description focuses on complementation. The analysis of findings demonstrates that, despite some similarities, almost all constructions with the two light verbs show either meaning or/and complementation differences. The study also reveals that both light verbs and deverbal nouns have an impact on the semantics and syntax of the construction. Light verbs may affect light verb constructions semantically in terms of generality and polysemy, and the aspectual meaning of constructions depends on both light verbs and deverbal nouns. Likewise, both light verbs and deverbal nouns have an influence on complementation, but in their own way: light verbs on the selection of complements and deverbal nouns on complementation patterns.


Introduction
Light verb constructions (LVCs) (e.g. have a chat, give a hug, make an attempt) have received considerable interest among scholars due to their complex nature and widespread use.The most recent studies analyse LVCs in terms of cross-linguistic features and translation strategies (Wang et al. 2023), morpho-syntactic features (Nugraha 2023), multiplicative meanings (Tron et al. 2022), transfer from one language to another (Barking et al. 2022) as well as distinctive structural and discursive features (Martínez Caro & Arús-Hita 2020).
The main feature of light verb structures is that they consist of a light verb (LV) the meaning of which is bleached and a nominal complement which provides the semantic content of the construction and is usually referred to as a deverbal noun (Huddleston & Pullum 2010;Plante 2014).Despite the weakened meaning, the light verb contributes to the semantics of the construction by adding some aspectuality, determining theta-role 1 assignment and valency of the construction as well as placing constraints on which complements can occur with particular light verbs (Wierzbicka 1982;Bergs 2005;Seiss 2009;Butt 2010).Light verbs may select a great variety of deverbal nouns and form semantic clusters (Bonial 2014).However, LVs themselves can show diversity in their use of deverbal nouns: there are numerous cases of the same deverbal nouns being selected by different light verbs (North 2005;Klich 2010).Due to these features of LVCs, the constructions pose difficulties to non-native language learners (Pathirana & Kumara 2011).
Constructions with the verbs do and make are among the most confusing in English.The two verbs have similar meanings, and deciding which should be used in a particular situation is sometimes difficult (Swan 1991;Allan 1998;Klich 2010).Some guidelines have been proposed to distinguish between the non-light uses of the verbs do and make when the meaning of the verbs is fully manifested as in make a cake and do a favour.The verb do is usually related to an action as a process and is used when talking about the action in general, whereas the verb make indicates "the end product of the action" (Carter & McCarthy 2007, 105) that results from creating or constructing something and refers to a specific action (Sinclair 1993;Tobin 1993, as cited in Allan 1998).However, according to Swan (1991), there are not always "clear rules" as to which of the two verbs should be used, and it is recommended to "look in a good dictionary or choose make -it is more likely to be right" (1991,180).These unclear cases are mainly related to the 'light' uses of do and make.
The English light verbs make and do are described as selecting a deverbal noun in the base (e.g.make a comment/promise/report; do a dance/jump/dive) or a derived form 1 Theta-role roles are semantic roles (e.g.AGENT, PATIENT) that are assigned by the predicate to its arguments (Cullicover 1997).
(e.g.make a statement/suggestion/decision; do a translation/cleaning/drawing) (Dixon 2005;Huddleston & Pullum 2010).According to Quirk et al. (1985), only the light do takes -ing deverbal nouns (Quirk et al. 1985).One more common feature characteristic of both verbs is that LVCs with make and do have telic meanings, which means that the deverbal nouns selected by both verbs express a goal-oriented and usually complete action (Bonial 2014;Plante 2014).Some scholars characterise the light verbs make and do as 'vague action verbs' as opposed to 'true light verbs' since, like ordinary transitives, they allow passivisation (e.g.An inspection was made some time last week.The ironing was done yesterday.),WH movement (e.g.Which offer did the finance company make?Which painting did John do?), pronominalisation (e.g.The Health Department made an inspection on Monday and may make another one before prosecuting.Remember, you did a drawing of that man last week.I think that you did one successfully. 2) and may select definite NP complements (e.g.I can't find the report and I don't know who made the inspection.These letters are nearly ready; I just have to do the typing.)(Kearns 2002, 21, 31).In the present investigation no differentiation between 'vague action verbs' and 'true light verbs' is made, and both types are treated as light verbs since they have the same structure and can be replaced by a verbal equivalent.
There have been a few debates on the semantic differences between the heavy and light uses of the verb make.When discussing the degree of acceptability of LVCs, Stevenson et al. (2004) state that it is difficult to differentiate between 'light' and 'heavy' uses of make (Stevenson et al. 2004).A similar idea is expressed by Allan (1998) who claims that the light verb make is very close in meaning to its heavy use as it is often related to creation or construction (Allan 1998).However, Kearns (2002) argues that it is possible to distinguish between light and heavy uses of make.She identifies two semantically distinct types of make depending on the deverbal noun: make with a bleached meaning when it is used with a deverbal noun denoting an action (e.g.make an inspection) and make expressing the "process of 'making' or 'creating' the referent" when combined with a deverbal noun that refers to a state or an abstract entity (e.g.make a noise, make a disturbance) (Kearns 2002, 23-25).Mehl (2019) supports Kearn's view by showing that the light verb make differs from the non-light in that it tends to resist coordinated direct objects (Mehl 2019).
There is a relatively small number of studies on the light verbs make and do when compared with other light verbs.Some research on the English light verbs make and/ or do has been carried out in terms of semantics (Bonial 2014;Plante 2014), syntax and grammaticalisation (Elenbaas 2011) and computational analysis (North 2005;Tu & Roth 2011).Both quantitative and qualitative investigations of the light verbs make and do have been conducted across different English varieties (Smith 2009;Laporte 2012;2 The pronominalisation of an LVC with the light verb do is illustrated by the author's example.Ronan & Schneider 2015;Mehl 2019;Sundquist 2020).LVCs with the light verb make have been analysed in terms of nativised structural patterns in Malaysian English (Ong & Rahim 2021).The light verb make has also been extensively described in a contrastive analysis in English and Norwegian (Klich 2010).However, generally, the two verbs have been studied along with other light verbs, and the analysis is rather fragmentary.
The comparison between the English light verbs make and do has also received very little attention.In these studies the verbs make and do are usually analysed along with other light verbs in terms of their frequency of use (Ronan & Schneider 2015), identification and extraction of LVCs from corpora (Tu & Roth 2011;Tan et al. 2021) and grammaticalisation (Elenbaas 2011).
The present investigation is a further attempt to compare LVCs with the light verbs make and do.The analysis will deal with the comparison of LVCs with the verbs make and do in combination with the same deverbal nouns to reveal similarities and differences between the use of the two light verbs in terms of semantic and syntactic features.To determine which part of LVCs is related to the similarities and differences, the contribution of the light verbs and deverbal nouns to the construction will be examined.Thus, the research will focus on the following questions: (1) What are the semantic and syntactic similarities and differences between the LVCs with the light verbs make and do in combination with the same deverbal nouns?(2) How do the light verbs and deverbal nouns contribute to the semantics and syntax of LVCs 3 ?
The aim of the study is to analyse light verb constructions with the verbs make and do when combined with the same deverbal nouns in terms of semantic and syntactic features on the basis of corpus data collected from the British National Corpus (BNC).
For this purpose, LVCs with make and do are examined in terms of lexical meaning and complementation 4 .

Related work
Numerous studies deal with the semantics of LVCs; however, only a few of them focus on semantic differences between LVCs with various light verbs and the contribution of light verbs to the semantics of LVCs.Some research on semantic differences between LVCs

3
Modification is not studied due to a very low number of modified LVCs with the light verbs make and do.

4
Complementation refers to complements selected by LVCs.It will be described in terms of complementation patterns which show the internal structure of complements of LVCs.
examines features of different light verbs.When analysing the idiosyncratic features of LVCs with have and take5 , Wierzbicka (1982) reveals systematic behaviour of LVCs and describes the actions expressed by LVCs with have as self-directed, whereas LVCs with take are associated with the actions starting at a definite moment and controllable (cf.have a walk vs. take a walk) (Wierzbicka 1982).Likewise, Dixon (2005) admits that LVCs with different light verbs add some "special semantic element to the basic sentence" (Dixon 2005, 463).When discussing the semantic conditions for LVCs, he shows that LVCs with the verbs have, give and take have both similarities and differences in terms of characteristic features they exhibit.LVCs with all three verbs express voluntary actions, but they differ in that the actions expressed by LVCs with have can be related to joy or relief, those with take involve some effort, while LVCs with the verb give express short duration single actions affecting the object in some way.According to Dixon, due to the semantic differences, the LVs tend to combine with different deverbal nouns (Dixon 2005).
Plante (2014) uses telicity (lexical aspect6 ) to show semantic differences between LVCs with different light verbs.When examining telicity of LVCs with the verbs have and do, Plante claims that the light verbs impose constraints on their complements: the verb have combines only with deverbal nouns that are atelic, whereas do with those that are telic.However, not all light verbs can be differentiated in this way: the LVs take and give do not seem to restrict their complements in terms of telicity (Plante 2014).The idea that LVs add some semantic information about the type of event is also supported by Seiss (2009) and Butt (2010).Butt states that LVs "signal some kind of boundedness or telicity or causation (crosslinguistically)" and admits that "the degree to which they signal this differs from language to language" (Butt 2010, 16).Bergs (2005) demonstrates that LVs can also contribute to the semantics of LVCs by assigning semantic roles to their constituents.For instance, in John did the operation and John had the operation, the subject in the former is an agent, whereas in the latter it is a patient.Thus, it follows that LVs are responsible for "the realization of particular argument control structures" (Bergs 2005, 215).However, Wittenberg and Piñango (2011) state that the deverbal noun may also be associated with semantic roles of LVCs.Comparing the non-light sentence Henry gives Elsa a rose with the sentence Henry gives Elsa an order containing a light verb construction, Wittenberg and Piñango show that in the latter the noun order is not only a carrier of a semantic role but also "licenses semantic roles to the other participants in the sentence" (Wittenberg & Piñango 2011, 393-394).
Most scholars admit that the deverbal noun is responsible for the lexical meaning of LVCs, while the meaning of the light verb is bleached (Bonial 2014;Plante 2014).In addition, the deverbal noun contributes to the grammatical aspect of LVCs (Wierzbicka 1982).When describing LVCs with have, Wierzbicka demonstrates that it is the deverbal noun that "delimits the duration of the action or event, and presents it as repeatable" (Wierzbicka 1982, 791).
The syntax of LVCs has been investigated by a number of scholars such as Kearns (2002), Dixon (2005) and Butt (2010) among others; however, complementation has been analysed only fragmentarily.Most of the research on the complementation of LVCs centres on the question of whether the elements following the deverbal noun are complements of the light verb or of the noun itself.The view that the complementation of LVCs is dependent on the light verb is supported by Wittenberg et al. (2014) and Bruening (2015).According to Bruening (2015), LVs determine the selection of arguments in LVCs since deverbal nouns do not retain the same arguments when combined with different light verbs: cf.She took a punch at him vs She gave him a punch (Bruening 2015, 6).When combined with the noun punch, the light verb take selects a prepositional phrase and does not allow an indirect object, whereas the verb give allows an indirect object but cannot be used with a prepositional phrase (ibid.).In the analysis of the argument structure of the light verbs give and take, Wittenberg et al. (2014) point out that both light verbs behave similarly to their heavy counterpart in the selection of arguments.For instance, the feature whereby the heavy verb give may select a double object and allow indirect object alternations is also characteristic of the light give: cf.give a hug to Harry and give Harry a hug.Similarly, the heavy take, being a transitive verb, takes a direct object functioning as the Theme, and in the LVC with the light take (e.g.take a walk), the deverbal noun is found in the direct object position, despite the fact that it is not the Theme but a part of a complex predicate (Wittenberg et al. 2014, 4).
The opposite view is represented by Huddleston and Pullum (2010) who argue that the elements following the deverbal noun are complements of the noun itself.Their main argument is that complements of the LVC can be selected by the noun "when it is used independently of the light verb": cf.His appeal for clemency failed and He made an appeal for clemency (2010,292).However, they admit that "properties of the light verb in its ordinary use may affect the complementation" (Huddleston & Pullum 2010, 293).For example, the verbs give, make and do are ditransitive in their heavy uses; however, when they are light, they can take indirect objects, and their selected deverbal nouns cannot (ibid.).
The present study will try to demonstrate to what extent the light verbs make and do contribute to the semantics and syntax of LVCs.

Data and methodology
The present study is corpus-based and mainly qualitative since it focused on the examination of concordance lines to identify semantic features and complementation patterns of the constructions under investigation.In some cases, the expanded context was used to decide on the meaning of constructions.In addition, it includes descriptive statistical analysis which is limited to the description of frequency of complementation patterns of LVCs with the verbs make and do and their distribution.
The research was carried out on the basis of two parameters: lexical meaning and complementation.The lexical meaning of LVCs was chosen for the analysis to identify whether the differences between heavy make and do were reflected in their light uses.The complementation of LVCs with make and do was studied to find out to what extent LVCs with the two light verbs differ syntactically.The two parameters were also employed to examine the impact of the light verbs and deverbal nouns on the semantics and syntax of LVCs.
The data for the analysis were collected from the British National Corpus (BNC), which is a representative electronic database of spoken and written British English.It contains 100 million words and includes spoken, fiction, newspaper, magazine, academic, nonacademic and miscellaneous registers.
In the data selection, the lemmatised form of the verbs make and do was tagged with all nouns ([make/do] a NOUN).Since LVCs are typically associated with constructions that have a verbal equivalent, only the structures that follow this requirement were selected for the analysis: e.g.make/do an evaluation have a verbal counterpart evaluate.Two lists of structures, one with the verb make and another with do, were compiled and juxtaposed to select the constructions where both verbs combine with the same deverbal nouns.The benefit of the choice to research LVCs with the light verbs make and do in combination with the same deverbal nouns is twofold.It allows a more elaborate study of similarities and differences between the two light verbs.In addition, it enables to examine the constructions when the behaviour of all LVCs with the verbs make and do is taken into account and when the verbs make and do in combination with particular deverbal nouns are compared, for instance, the use of make and do with the deverbal noun survey.
Next, to include occurrences with different determiners and internal modifiers, each of the selected structures with the verbs make and do that have a verbal equivalent was searched with a span of three words between the verb and the nominal complement (inserting three asterisks in between): e.g.[make/do] *** evaluation.Thus, the data selected for the analysis contained not only the cases where the deverbal noun was used with the indefinite article, which is one of the characteristic features of LVCs, but also when it was found with the definite and zero articles, as well as other determiners: e.g.It is actually easier to do the calculation if instead of individual turns we assume that the current is continuously distributed […].;I am still to be convinced whether the energy and time required to do effective evaluation would balance the increased efficiency in teaching which might follow.; […] he had already made his visit to the hospital to see how Byford was getting on.Before the final list was compiled, the concordance lines with the selected strings were studied in search of idiomatic constructions.The idioms 'make a/the calculation' meaning 'guess something using facts' (e.g.Coleridge himself cannot have failed to make the calculation that of four marriageable Fricker sisters, one was married to Robert Lovell […].), 'make a presentation' meaning 'give a prize' (e.g.The SFA will, 20 minutes before todayʼs kickoff, make a presentation to those players who were part of that commendable run.)among others were excluded from the research 7 .Since context is essential for semantic analysis, the structures under investigation used without any context (titles and subtitles) were also eliminated.
The number of occurrences was also an important criterion in the data selection.The constructions that occurred at least five times in the corpus were selected for the analysis.However, there were two exceptions made: the LVCs make a painting and do an examination that had only four occurrences were also chosen for the study for comparative purposes as the former is synonymous to make a drawing and the latter to make a check.
The total number of LVCs selected for research was 342: 180 with the verb make and 162 with the verb do.As seen from All in all, 22 constructions with the verb make and six with the verb do were discarded.
The analysis of the light verb constructions under investigation was carried out in the following way.First, the meaning of all the collected LVCs with make and do was identified by examining concordance lines and using the online Cambridge Dictionary (2022).
Then, the LVCs with the verbs make and do were juxtaposed and grouped according to similarities and differences in meaning.Next, all the groups of LVCs were described providing examples to illustrate different peculiarities of the constructions.Further study focused on syntactic features of LVCs with make and do.The description of syntactic features included complementation frequency and patterns as well as the distribution of complements across different deverbal nouns.Due to a small number of complemented cases, two types of filters were used when selecting complements.The complementation analysis included the complements that occurred at least three times and those that made up at least 50% of all occurrences of LVCs.The selected complements were organised into groups according to complementation patterns.Then, the distribution of the patterns across different LVCs was studied, the frequency was calculated and the complementation of LVCs with make and do was analysed and described.Finally, the contribution of light verbs and deverbal nouns to the semantics and syntax of LVCs is discussed.

Findings and discussion
In this section, a comparative analysis of LVCs with the verbs make and do when combined with the same deverbal nouns is carried out in terms of semantic and syntactic features.Semantically, three types of features are described: generality of meaning 8 , the polysemy of deverbal nouns and aspectual meaning.The syntactic analysis includes the complementation frequency and patterns as well as their distribution across LVCs with the light verbs make and do.

Description of semantic features
The semantic analysis shows that LVCs with the verbs make and do can be divided into five groups: LVCs with make and do that have the same meaning and those that have some difference in lexical meaning, LVCs where the use of make and do is related to a different degree of generality, LVCs where the choice between the two verbs leads to a wider range of meanings, i.e. polysemy 9 , and LVCs where the use of make and do is associated with aspectual meaning differences.
The LVCs with make and do where no semantic difference is observed include four deverbal nouns out of twelve: copy (1a), examination (1b), check (1c) and visit (1d). 8 Generality of meaning can be understood as being opposite to specificity of meaning, i.e. it refers to cases when the context in which the construction is used is not specified.9 Lexical meaning and polysemy are closely related notions.In the present study, when describing semantic differences, the former refers to the use of different meanings, and the latter to differences in the variety of meanings.
(1) a. […] he can make a copy of a hyperdocument and edit the context-specifications within the context- Examples 1a and 1b above show that the indication of the object undergoing the action expressed by the light verb construction is associated with the use of the light verb make since the LVCs with make are typically followed by a complement, whereas in cases of LVCs with do the object is usually identified from the context.Thus, what is copied in (1a) is "files", and what is examined in (1b) is "greenbelt boundaries".Similarly, when the noun visit is used with the verb make, the place of the visit is typically specified (1d), but the indication of the place when visit combines with do can be deduced only from the wider context.In example 1c both make and do in combination with check require the context for the object undergoing the action to be identified.
The object that undergoes the action expressed by the light verb construction sometimes cannot be understood even from the context.This absence of the object can account for the different degree of generality of LVCs.The difference is characteristic of LVCs when make and do combine with the deverbal nouns evaluation, presentation, drawing, painting and sketch.These deverbal nouns in combination with do (2 a-e) tend to have a more general meaning than when combined with the verb make (2 f-j).This does not mean that the object cannot be found with constructions with the light verb do at all; however, such cases are rare.For instance, the object is present when the verb do combines with the deverbal noun drawing in two cases out of 51 and with the noun presentation in one case out of 10.
(2) a.In my mind, the object of doing an evaluation is to create some kind of improvement in the situation that is being evaluated.e.The attraction of doing a sketch like this is the speed at which it can be achieved.
f.I wanted to make an evaluation of his condition after his appearance against Tottenham on Wednesday.
g.The generality of LVCs with drawing and painting in combination with the verb do may, in addition, be emphasised by the use of determiners expressed by indefinite pronouns (3a, b) and modifiers manifested by classifying adjectives (3c, d).
(3) a. Tish intends to do some drawing, but is too intent on keeping a diary.
b.You can do some painting when you go to nursery can't you? c.They acquire multiple degrees, purposelessly log facts, or do photoreal drawing […]. d.
[…] they all do oil painting and the music they like is shite.
Cambridge Dictionary (2022) demonstrates that all deverbal nouns under investigation, except for the noun evaluation, are polysemous; however, most of them in combination with the light verbs make and do are used in one of the meanings.The difference in the variety of meanings between LVCs with the verbs make and do can be traced only when they combine with the deverbal nouns painting, survey and tour.The noun painting in combination with the verb make has the meaning of making a picture using paint (4a, b), but when combined with do it may have an additional meaning of putting paint on a wall (4c) (CD 2022).In a similar way, when the light make combines with the noun survey, it usually has the meaning of studying something carefully in order to form some opinion (4d, e), whereas in combination with the verb do it has, in addition, the meaning of asking people questions in order to find out their opinion or behavior (4f) (ibid.).
(4) a.There had been the time, too, when Ruth made a painting of a dragon devouring a knight which had apparently scared another child.

b. Helen may do some painting and Elizabeth likes to work at drawing shapes
c. How many Swedes does it take to paint a wall? A. Twenty-seven.One to do the painting and twenty-six to organize the spectators. d.
[…] he proposed that he and Moray make a survey, by boat, to consider if anything, might be done protectively.
e.They were discovered by environmentalists doing a survey of the region's flora and fauna.
f. Before the election, World In Action did a survey of 250 unregistered people in Hampstead and Highgate.
Likewise, the combination of the verb do with the noun tour is polysemous as it may have not only the meaning of going to a place or area (5b), which is also typical of the construction make a tour (5a), but the meaning of going on a planned visit for a special purpose (5c) as well.
( The combination of the deverbal noun calculation with the light verbs make and do results in LVCs not only with a slightly different lexical meaning but also the difference in lexical aspect.When calculation combines with make, LVCs have the meaning of making a decision on the basis of the calculation (6a), whereas the construction do a calculation refers to the process of calculating itself (6b).Thus, the construction make a calculation is related to result and do a calculation to process.
(6) a.The cashier grumbled about changing Swiss money, but made a calculation that was to his advantage on the rate.

b. It is actually easier to do the calculation if instead of individual turns we assume that the current is continuously distributed on the surface of the cylinder […].
The deverbal nouns evaluation (7), drawing (8), painting (9) and survey (10) also tend to show aspectual differences in combination with make and do.In some cases, the aspectual meaning of LVCs with these deverbal nouns is not clear-cut due to the context in which they are used: the activity in some way is merged with the accomplishment.Therefore, a sub-type of the event which is a combination of process and result is introduced.In such instances, the aspectual meaning can be affected by the mention of the purpose of the activity or the absence of the goal of the activity.When used with the verb make, the constructions can refer to result (7a, 10a) or a combination of process and result (8a, 9a), and in combination with the verb do, they may be related to process (7b, 8b, 9b, 10b) or a combination of process and result (7c, 10c).
(7) a.This is not without problems, such as when to intervene and make an evaluation.
b. Now my feeling is that in doing an evaluation one ought to try and develop each of these different viewpoints […].
c.I am still to be convinced whether the energy and time required to do effective evaluation would balance the increased efficiency in teaching which might follow.
(8) a.She was making a drawing of a tall woman in flowing sleeves.
b. Would you like to go and get a card out of there look, and do some drawing?
(9) a. Honor West sat in her bedroom making a painting of the roses which grew just beyond her window. b.
[…] perhaps you can come and do some painting with the residents.c.We now have to do a survey to find out if there was a late swing.
Despite the fact that LVCs with both make and do may be associated with the aspectual meaning which is the combination of process and result, the overall tendency of the verb make to be related to result and do with process can be observed.
Other deverbal nouns when combined with make and do show no aspectual differences.
For instance, both make and do a presentation (11a) refer to result, make and do a copy (11b) to the combination of process and result, and make and do an examination (11c) and make and do a tour (11d) to process.
(11) a.To sum up, the findings demonstrate that some LVCs with make and do are similar in that they may have the same meaning.The group includes LVCs with the deverbal nouns copy, examination, check and visit.Semantic differences between LVCs with make and do are related to generality, polysemy, lexical meaning and aspect.The difference in generality of meaning applies to LVCs with the deverbal nouns evaluation, presentation, drawing, painting and sketch.The general meaning of such LVCs is associated with the verb do and the specific meaning with the verb make.The polysemy of LVCs with the deverbal nouns painting, survey and tour is also linked to the light verb do.There is only one case when the LVCs under investigation exhibit lexical meaning difference: it is when the verbs make and do combine with the noun calculation.Aspectual meaning differences are found when the verbs make and do combine with the deverbal nouns calculation, evaluation, drawing, painting and survey.The resultative meaning is usually associated with the verb make, whereas the processual meaning relates to the verb do.

Description of syntactic features
The investigation of complementation includes the frequency and variety of the complementation patterns of the LVCs with make and do.The frequency analysis includes the description of the total frequency of complements selected by LVCs with the verbs make and do and the frequency and distribution of the complements across different deverbal nouns.The total frequency of complements is counted by comparing the total number of complements selected by LVCs with the verbs make and do and the total number of LVCs found with each verb.Similarly, the frequency of complements with each deverbal noun is counted by comparing the number of complements of LVCs with make and do and the total number of LVCs with each of the verbs.
Table 2 shows that LVCs with the verb make select complements almost three times more often than with the verb do, 39% and 14% respectively, though the number of LVCs with make is only slightly greater than of those with do (180 with the verb make and 162 with the verb do).The number of deverbal nouns in combination with make and do that select complements differs significantly as well, 11 and three respectively.Calculation is the only deverbal noun that takes no complements in combination with both verbs.The deverbal noun check is the only deverbal noun that has no complements when combined with make.Two deverbal nouns survey and tour take complements when combined with both verbs; however, with the verb make, they select complements more than twice as often as with the verb do.As seen from Table 2, the distribution of complements across LVCs with different deverbal nouns differs greatly.Depending on the deverbal noun, the frequency of LVCs with the verb make ranges from 16% to 89% and with the verb do from 11% to 43%.The deverbal nouns drawing, copy and tour have the greatest number of complements when combined with the verb make, 89%, 72% and 72% respectively, whereas the largest number of complements of the verb do is found with the deverbal nouns check and tour, 43% and 32% respectively.
Figure 1 shows that the LVCs under investigation take four complementation patterns, and only one of them, namely, of + NP is common for both verbs.It is evident that LVCs with make select a greater variety of patterns than those with the verb do.(12) a.She was making a drawing of a tall woman in flowing sleeves. b.
[…] make some evaluation of the effectiveness of the GRIDS method after about twelve months.The indirect object complementation pattern to + NP is found only when make combines with the noun presentation (13a, b).
(13) a.I am asked to make a presentation to the Committee which was set up by the minister of finance. b.
[…] the head of maths did a presentation to the whole school […].
The complementation pattern of NP + to NP selected by LVCs with make could be related to a ditransitive structure.It is also limited to the construction make a presentation (14a, b). ( 14) a. Police are preparing to make a presentation of their proposal to councillors and town hall officials in the near future.b.Mr Hayton said he would be willing to make a presentation of the computer findings to councillors […].
LVCs with the verb do take two complementation patterns: of + NP and on + NP.Both patterns are characteristic of the verb do in combination with the deverbal noun survey (15a, b, d).In addition, the pattern of + NP is found with the noun tour (15c), and on + NP with the noun check (15e).It should be noted that complements could be found with both meanings of the noun survey: (1) studying something carefully in order to form some opinion (12a, d) and (2) asking people questions in order to find out their opinion or behavior (12b).However, both patterns are characteristic of LVCs with survey only with the first meaning.
All in all, the findings on complementation show that LVCs with only two deverbal nouns calculation and tour have no complementation differences when combined with make and do.Complementation differences between LVCs with the verbs make and do include the presence of complements as well as the variety of complementation patterns.The former difference can be found with LVCs with almost all deverbal nouns, except for survey, check and tour which are the only ones to take complements when combined with the verb do.LVCs with the verbs make and do also differ in the variety of complementation patterns they select.The prevailing pattern of LVCs with both verbs is of + NP; however, LVCs with make additionally take the patterns of NP + to NP and to + NP, and LVCs with the verb do can also select the pattern on + NP.

Summary of findings. Contribution to semantics and syntax of LVCs
The analysis of findings demonstrates that there are both similarities and differences between LVCs with the verbs make and do in terms of semantics and complementation.
As shown in Table 3, LVCs with the deverbal nouns copy, examination, check and visit when combined with the verbs make and do convey the same meaning.Table 3 also demonstrates that LVCs with two deverbal nouns calculation and tour have no complementation differences when combined with make and do.The noun tour in combination with the two verbs selects complements of the same pattern, whereas LVCs with the noun calculation do not take complements in either case.The similarities related to semantics and complementation include different deverbal nouns, which signifies that these LVCs differ if not semantically then syntactically, and this is how LVCs with make can be differentiated from those with the verb do.
*The difference refers to the indication of the place.
Table 3. Summary of similarities and differences of LVCs with the light verbs make and do Semantic differences between LVCs with make and do are related to generality, polysemy, lexical meaning and aspect.Table 3 demonstrates that eight out of 12 deverbal nouns that combine with both light verbs have semantic differences of some type.The deverbal nouns evaluation, drawing, calculation and survey can be related to two types of semantic difference, and the noun painting to three types of difference out of four.
As can be seen from Table 3, the difference in generality of meaning applies to LVCs with the deverbal nouns evaluation, presentation, drawing, painting and sketch.It should be noted that the generality of meaning in these cases can be related to the absence of complements; therefore, the interconnection between semantic and syntactic differences can be traced in this case.The general meaning of such LVCs is associated with the verb do and the specific meaning with the verb make, which shows the contribution of the light verbs to this particular aspect of the meaning of LVCs.One more contribution of make and do to the semantics of LVCs can be attributed to the ability of the light verbs to place constraints on the meaning of polysemous deverbal nouns they combine with.The polysemy of the deverbal nouns painting, survey and tour is evident when they combine with the light verb do, whereas in combination with make only one meaning of the deverbal nouns is used.Generality, combined with the polysemy of the deverbal nouns in combination with the verb do, suggests that semantically the light verb do is more universal than the verb make.The lexical difference can be associated with only one deverbal noun, namely, calculation which seems to show only semantic differences when combined with make and do.
The dependence of aspectual meaning on the light verb can be related to the deverbal noun calculation, evaluation, drawing, painting and survey.LVCs with the verb make tend to express the resultative meaning, whereas those with the verb do the processual meaning.All four types of semantic differences added together show that the light verbs have a rather great impact on the semantics of the construction.This is in line with previous research by Seiss (2009), Butt (2010) and Plante (2014).However, deverbal nouns are not devoid of contribution to semantics either.There are a few reasons for that.There are a number of cases where LVCs with make and do exhibit no differences in terms of lexical meaning, polysemy, generality or aspectual meaning.A variety of semantic differences as such could also support the idea.
Complementation differences between LVCs with the verbs make and do include the presence and frequency of complements as well as the variety of complementation patterns and their distribution across different deverbal nouns.All the differences show that complementation is far more characteristic of LVCs with the verb make than with do.
Table 3 shows that the most frequent complementation difference is related to the presence of complements.LVCs with the verb make select complements in combination with even 11 deverbal nouns out of 12, whereas those with the verb do take complements when combined with only three deverbal nouns.The fact that in most cases only LVCs with the verb make tend to take complements indicates that the selection of complements seems to be restricted by light verbs.This finding supports the claim that LVs determine the selection of arguments in LVCs in previous studies by Wittenberg et al. (2014) and Bruening (2015).
LVCs with the verbs make and do also differ in their selection of complementation patterns.The prevailing pattern of LVCs with both verbs is of + NP; however, LVCs with make additionally take the patterns of NP + to NP and to + NP, and LVCs with the verb do can also select the pattern on + NP.LVCs with the deverbal noun survey in combination with do select more complementation patterns than in combination with make: the patterns of + NP and on + NP with the verb do and of + NP with make.This might be an indication that light verbs also affect the complementation of LVCs in terms of patterning.However, the difference is limited to only one deverbal noun; therefore, further investigation is needed to confirm the statement.
The influence of deverbal nouns on complementation patterning seems more prominent as it can be revealed in two ways.One of them is related to the variety of patterns selected by LVCs with make and do.As mentioned above, LVCs with make may select three different patterns, and those with do two patterns.Only one pattern, namely, of + NP, is common for both.Another way that shows the dependence of patterning on deverbal nouns is the very structure of patterns.The most frequent complementation pattern of + NP found with both light verbs cannot be selected by verbal equivalents of LVCs, but can be used with the deverbal nouns (e.g. the evaluation of effectiveness, a drawing of a young lady, an examination of the concept of personal autonomy (BNC)).Two more patterns of NP + to NP and to + NP found when the light verb make combines with noun presentation can be selected by neither corresponding verbs of LVCs nor deverbal nouns, which means that the selection of the patterns could be possible only by combining the light verb with the deverbal noun.
The present research has some limitations that should be taken into account.One of the limitations is a lack of previous studies related to a more elaborate description of semantics and syntax of the light verbs make and do as well as the comparison of both light verbs, and most research on semantics and syntax of light verbs is outdated.The size of the collected data and the small number of complements of LVCs with some deverbal nouns might also be considered shortcomings of the study.

Concluding remarks
The research shows that the light verbs make and do in combination with the same deverbal nouns can be differentiated on the basis of semantic and syntactic features.There are a bunch of features rather than some distinct feature enabling such a differentiation.Meaning differences between LVCs with make and do are found with about 67% of the deverbal nouns (eight out of 12), whereas complementation differences are related to 83% of the deverbal nouns under investigation (10 out of 12).However, the combination of meaning and complementation as a means for distinguishing between LVCs with make and do is effective for almost 92% of deverbal nouns (11 out of 12).The remaining deverbal noun does not take complements, but, even in this case, the difference between the verbs make and do can be related to indication of place.
The comparison of LVCs with the verbs make and do also reveals that both light verbs and deverbal nouns may contribute to the semantics and syntax of LVCs, and the contribution shows that, to some extent, the light verbs make and do resemble their heavy counterparts.
It is widely accepted (Huddleston & Pullum 2010;Plante 2014;Ong & Rahim 2021) that the lexical meaning of LVCs depends on deverbal nouns, and this study suggests that the lexical meaning can be affected by light verbs in terms of generality, which is characteristic of LVCs with the verb do.In addition, light verbs may place constraints on which meaning/s polysemous deverbal nouns can have in the formation of LVCs.
According to previous research (Seiss 2009;Butt 2010;Plante 2014), light verbs are also responsible for the aspectual meaning of LVCs.The analysis of LVCs make and do confirms this statement since in some cases LVCs with the two verbs exhibit differences in terms of aspectual meaning.The study also shows that the aspectual meaning of make and do in LVCs may differ from the aspectual meaning characteristic of heavy uses of both verbs.The verb make, which typically has a resultative meaning in its heavy uses, may be associated with process or a combination of process and result when used in LVCs.Likewise, the processual meaning of the verb do in its non-light uses can change into resultative or a combination of process and result in LVCs.The research reveals an interesting fact that deverbal nouns may also have an impact on the aspectual meaning of LVCs with the verbs make and do since there are cases when they show no aspectual difference at all.
The study also supports the existing opposite views on the complementation of LVCs demonstrating that both light verbs and deverbal nouns can have an influence on the complementation of LVCs (Huddleston & Pullum 2010;Wittenberg et al. 2014;Bruening 2015), but in their own way.The analysis of LVCs with make and do shows that light verbs have an impact on the selection of complements, whereas deverbal nouns on complementation patterns.
The results are based on the investigation of only deverbal nouns combining with both make and do, and the number of LVCs with some deverbal nouns is rather small, as is the number of some complementation patterns.For this reason, further analysis including a greater variety of light verbs and deverbal nouns is needed.Further research could also include a comparison of light verb constructions make and do with their verbal equivalents as well as a study of modification features.

b.
At one TV training session I swapped the whole room around, putting the audience on the stage and doing the presentation from the floor!c. […] there's no reason why you shouldn't draw, do a drawing and check the drawing yourselves.d.For this I decided to do a painting looking from my conservatory through the windows into the garden […].
) a. Li Peng paid a visit to North Korea in May 1991 and made a tour of South-East Asia in August 1990 and of the Middle East in July 1991.b.Lorna decided to do the tour of South Pacific.c.Producer Duncan Weldon has asked him from time to time to do a tour butCourtenay has never responded until now.

(
10) a. Brooke and Remmers (1970) made a survey of 115 foreign subsidiaries operating in the UK […].b.The girl comes round here doing a survey, she came round one day and we were civil to her […].

Figure 1 .
Figure 1.The variety and distribution of complements of LVCs with make and do c. […] a party may inspect and make a copy of any document referred to on the Land Register […].d.We were very glad to arrange for five photographers to make a survey of the village; its people and occupations […].e.I recently made a tour of India which was arranged by ANZ Grindlays Bank […].

(
15) a. […] by the time the DET did a survey of farm schooling in 1986, only 3 per cent of the total of 5,782 farm schools offered education for more than seven grades […].b. […] World In Action did a survey of 250 unregistered people in Hampstead and Highgate.c.Having narrowed the field, do a tour of the shops and try as many different models as you can […].d.I did a survey on the most popular pop star and those two came out tops in the 'best-looking' stakes.e. […] we're going to employ a consultant […] to do a check on a bridge.

Table 1 .
Table 1, the selected constructions included LVCs with twelve deverbal nouns: evaluation, presentation, copy, drawing, painting, sketch, calculation, examination, survey, check, tour and visit.LVCs with make and do selected for research 7

table .
And if you've got the same file names, and you're doing a copy in DOS, you just overwrite them […].
d. Board has just made a visit to the North West Region […].That reminds me of the time at our branch they're doing a visit, and they didn't come down my isle […].
The British Consul General Dennis Doble made the presentation of Wedgwood plaques[…].
h.In 1967, my wife Ann made a drawing of Stenness, including a stone bench across two stubby, central columns.i. Honor West sat in her bedroom making a painting of the roses which grew just beyond her window.j.Viktor Rakovsky had made a sketch of the scene; he would finish it later.

Table 2 .
Variety and frequency of complementation patterns of LVCs with make and do and their distribution