Raymond Ip

76-451 Patterns of English Usage

Paper for Term Project

Shall we say “We Will” or Will we say “We Shall”?

May 1, 2000

Introduction:

As presented in the American Heritage Dictionary of English Usage (1996), traditional rules for shall and will usage are as follows: (1) Simple futurity can be expressed by using “shall+infinitive” in the first person and “will+infinitive” in the second and third persons; (2) To convey a sense of determination, promise, or obligation, one should use “will+infinitive” in the first person and “shall+infinitive” in the other persons. However, the dictionary has also indicated that the aforementioned prescriptions for shall and will have “never taken firm root outside of … the English of the English (as opposed to that of the Scots and Irish)” and are “subject to variation” even in British English. In fact, the dictionary’s editors advise people not to use “shall” if they are uncertain, so as to avoid the “considerable risk of getting it wrong”. Indications that the traditional rules may be outdated can also be found in the text by Quirk and Greenbaum (1973: 47-48), who states that will is applicable to not only the second and third persons, but also the first person.

Currently, there are widely held suppositions that (1) will has pretty much replaced shall in most instances of usage and (2) shall is gradually fading out (except when it is used to convey emphasis and / or resolution in the second and third persons). My suspicion is that these common suppositions only apply when shall and will are in their sentence form[i] (e.g., “I shall”, “you will”, and “they shall”) and that a different pattern may exist for the question form of shall and will (i.e., “shall I”, “will you”, “shall they”, etc.)

Main Objectives:

The main goal of this project is to utilize the Cobuild Corpus in evaluating the validity of the following published or widely held contentions:

(1) Traditional rules for shall and will usage are outdated;

(2) Such rules are followed more closely in British than in American English;

(3) Will is replacing shall in most instances of usage (for both sentence and question forms).

Another goal of the project is to examine the overall distribution of shall and will usage. Brief investigations will be conducted to look at the potential differences in usage patterns (1) between spoken and written English, (2) among the different persons (first, second, and third) and number (singular and plural).

Procedures, Results, and Discussion:

(1) Are the traditional rules for shall and will usage outdated?

(1-a) Procedures:

The following queries were entered into Cobuild: <i+shall>, <i+will>, <we+shall>, <we+will>, <you+shall>, <you+will>, <he|she|it+shall>, <he|she|it+will>, <they+shall>, <they+will>. The entire corpus is used. Then, irrelevant instances from the output were edited [ii] (For instance, in the “you+shall” output, a line like “I love you. Shall we get married?” will be eliminated). Afterwards, the total number of relevant instances of shall and will were counted for each of the personal pronouns.[iii] Lastly, “Will / Shall” ratios (abbreviated as WSRs from this point on) were calculated using the following formula:

WSR = (Instances of will) / (Instances of shall)

Eqn. 1

(1-b) Results and Discussion:

Table 1: Variation of “Will / Shall” Ratios Across Different Persons and Number – Sentence Form [iv]

“Will / Shall” Ratio (= Instances of “Will” / Instances of “Shall”)

1^st Person Singular

“I”

1^st Person Plural

“We”

2^nd Person

Singular/Plural

“You”

3^rd Person Singular,

“He / She / It”

3^rd Person Plural

“They”

Overall

4.3

6.5

70.8

197.0

142.6

As seen in Table 1, the WSRs for all five cases (1^st person singular and plural, 2^nd person combined, 3^rd person singular and plural) are larger than one. In other words, will is used more frequently than shall regardless of the person of the pronoun. This observation is an apparent contradiction of the traditional grammar rules, which state that shall is the correct auxiliary verb to use when first person is involved. According to Jespersen (1964:280-281), this preference for will over shall may be attributable to phonetic ease. He comments that “while we have no examples of the dropping of the sound [∫], the sound [w] often disappears in weak positions”. Examples he gives include words like answer, Greenwich, and Southwark. What Jespersen is suggesting is that expressions like “you will” or “I will” can be pronounced quickly (as “you’ll” or “I’ll”), without major alterations from how the original expressions sound. On the other hand, there is no quick way of articulating “you shall” or “I shall” without significantly changing how they sound.

However, even though will is generally preferred over shall, the degree of preference varies among the three persons. The preference for will is the strongest when third person is involved – the range of WSRs is around 140-190. Third person pronouns are followed by their second person counterparts, which have a combined WSR of around 70. The preference for will is the weakest in the first person (the two WSRs are both less than 10). One possible explanation for these observations is that shall and will have fundamental differences in their meanings across different persons (For simplicity, in later discussions, I will refer to this line of reasoning as the “Variation in Meaning” Hypothesis). In second and third persons, shall is used almost exclusively in cases when there is a sense of emphasis or intensity. Examples are orders, resolutions, rules, and so forth:

Emphasis:	You’ll watch it, I swear you shall!
Order:	God d--- you, you shall* not go, for we allow to kill you.*

There are very few instances in second and third persons when shall is used to indicate simple futurity. On the other hand, will can both indicate simple futurity and convey a sense of emphasis. In other words, will has a broader meaning than shall in the second and third persons. On the contrary, the line of demarcation between will and shall in first person is not as clear-cut – they can both convey emphasis and simple futurity. In cases when the writer (speaker) chooses to indicate the degree of emphasis through the tone of the prose (speech), will and shall are almost interchangeable. In other words, there are fewer (and not-as-strong) reasons to choose will over shall in the first person. This may have explained why the WSRs are substantially smaller for the first persons.

In addition to explaining the difference between the first person and the other two persons, the “Variation in Meaning” Hypothesis also accounts for the different WSRs for the second and third persons – the second-person shall seems to have broader applications than the third-person shall. As mentioned before, shall is generally used for emphasis in the second and third persons. Since rules or commands (directed more often to the second than to the third person) constitute a large portion of these cases of emphasis, there may be more instances of shall (relative to will) in the second than in the third person. This explains why the second-person WSR is smaller than that of the third person.

Yet, still one question remains: “Are there other possible reasons why the first-person WSRs are much smaller than their second- and third-person counterparts?” Perhaps, the predominant trend is that for first person, will and shall are to be used in the same fashion as they are in the other two persons (i.e., shall is used in cases of emphasis but not in cases of simple futurity, whereas will is used in both cases). However, the traditional grammar rules, which suggest the exact opposite pattern of usage,[v] may still have some “residual” influence on how people use the two auxiliary verbs.[vi] The counteraction of the new trend against the old rule may have resulted in the “compromised” and smaller WSR for the first-person cases.

(1-c) Conclusion:

Looking at the data (will is used four to six times more often than shall in the first person), we can see that the traditional rules for shall and will usage are somewhat outdated, at least for the first person.[vii] However, as discussed earlier, the preference of will over shall is significantly weaker in the first person than in the other two persons. This observation is not entirely consistent with the “new rules” of shall and will usage. One implication may be that the traditional rules, although not as prevalent as before, may still be at work on some level and have resulted in the weaker preference of will over shall. Another possible explanation is that shall and will have more similar meanings in the first person and thus people have a more difficult time in choosing between the two (versus in second and third persons, where shall are mostly restricted to cases of emphasis).

(2) Are the traditional rules more closely followed in British than in American English?

(2-a) Procedures:

To compare patterns in American versus British English, we used only six (three pairs) of the corpora in Cobuild: the US and UK ephemeral (usephem, ukephem), the US and UK books (usbooks, ukbooks), and the two radio stations (npr, bbc).[viii] We began by putting query lines (same as those in part 1) into each of the six corpora separately. We then calculated the WSRs for each corpus using the same method as discussed in part 1. Lastly, we determined the US/UK ratios (abbreviated as USUKRs) for each of the three pairs of corpora using the following equation:

USUKR = (WSR for an American corpus) / (WSR for its British counterpart)

Eqn. 2

(2-b) Results and Discussion:

Table 2: “US / UK” Ratios for Different Persons / Number & Across Different Corpora – Sentence Form

“US / UK” Ratio [ = (US “Will / Shall” Ratio) / (UK “Will / Shall” Ratio) ]

1^st Person Singular

“I”

1^st Person Plural

“We”

2^nd Person

Singular/Plural

“You”

3^rd Person Singular,

“He / She / It”

3^rd Person Plural

“They”

Overall

2.7

1.1

0.031

0.063

0.22

Ephemeral

13.0

25.8

1.3

0.87

CBD [ix]

Books

2.5

1.54

0.21

0.12

0.13

Radio

0.96

2.83

As reported in Table 2, for second and third persons, the USUKRs are consistently smaller than (or close to) one. In other words, the British WSRs are generally larger than the American WSRs for these two persons. There are several possible explanations for this observation. One is that the British adheres more closely to the grammatical rules of using will over shall. This leads to a larger British WSR relative to the American WSR, which in turn results in a smaller USUKR. Another possible explanation is that Americans use the emphatic tone more often (which implies the more frequent usage of shall for second and third persons). The increased usage of shall relative to will then results in a smaller American WSR. This can ultimately render a smaller USUKR in second and third persons.

The first person demonstrates a fairly different pattern from that of the second and third persons. Unlike the other two persons (both with USUKRs clearly smaller than 1), the first person has USUKRs larger than or close to one (singular USUKR of 2.7 and plural USUKR of 1.1). This suggests that British speakers do have a slightly higher preference for shall than the Americans – relatively speaking [x].

One other point worth noticing is that the three pairs of corpora give drastically different USUKRs from each other for cases involving first person (see Table 2). In other words, the first person USUKR shows little consistency across corpora. For instance, for first person singular, the ephemerals have a USUKR of 13.8 – fourteen times larger than the USUKR of the “radio” category.

(2-c) Conclusions:

As mentioned earlier, there is some evidence that (1) in cases involving first person, the British has a slightly higher preference for shall than the Americans; (2) in cases of second and third persons, the British has a slightly higher preference for will. Again, these comments about the British’s preferences are all relative (see endnote x). In other words, the evidence we have is by no means conclusive and can neither confirm nor deny the widely held contention that traditional rules for shall and will are followed more closely in British than in America English.

(3) Is the replacement of shall by will only occurring when shall and will are in sentence form but not when they are in question form?

By far, we have restricted our analysis to sentence form (i.e., “pronoun+shall/will”) and have not paid any close attention to the question form (i.e., “shall/will+pronoun”). In fact, when most people consider the usage patterns of shall and will, they almost always focus on the sentence forms exclusively. My hypothesis is that the question form may have its own pattern of shall and will usage.

(3-a) Procedures:

The steps were identical to those presented in section (1-a). The only differences were in the query lines used: <shall+i>, <will+i>, <shall+we>, <will+we>, <shall+you>, <will+you>, <shall+he|she|it>, <will+he|she|it>, <shall+they>, <will+they>. The WSRs for the question form and those for the sentence form were then calculated and contrasted.

(3-b) Results and Discussion:

Table 3: Comparison of “Will / Shall” Ratios for Sentence versus Question Forms

“Will / Shall” Ratio (= Instances of “Will” / Instances of “Shall”)

1^st Person Singular

“I”

1^st Person Plural

“We”

2^nd Person

Singular/Plural

“You”

3^rd Person Singular,

“He / She / It”

3^rd Person Plural

“They”

Question

0.57

0.24

81.9

51.0

52.7

Sentence

4.3

6.5

70.8

197.0

142.6

The most interesting observation in Table 3 is that shall is actually more prevalent than will in the question form of the first person (WSRs of 0.57 and 0.24 for singular and plural respectively). This finding is most certainly against the widely held beliefs on how shall and will are used in first person. Similar to the observations for the sentence form, the second and third persons also have much larger WSRs than the first person in question form. However, for the third person, the question-form WSRs were only about one-third of the sentence-form WSRs.

So, what are some possible reasons why “shall I/we” is more prevalent than “will I/we”? One plausible explanation is again the “Variation in Meaning” Hypothesis. In first-person sentence form, shall and will have similar meanings and are interchangeable in many cases. However, in question form, shall seems to carry broader meanings and functions relative to will. First of all, other than indicating simple futurity, “shall I/we” can also be used as a polite way of asking a question. For instance,

Shall I get you something? (c.f., Will I get you something?)

“Shall I/we …” can also be used to convey the meaning of “Let me / us …”.

Shall we go around… Start(ing) with you. (c.f., Let’s go around… Start(ing) with you)

In fact, expressions like “shall we say” or “shall I say” are almost like idiomatic expressions – approximately 16% of the sample outputs from the query <shall+we> are “shall we say”. These additional meanings that shall has in the first-person question form may have contributed to the WSRs of less than one.

Diagram 4: Illustration of the Contraction Hypothesis

In their Sound like People begin to use will not /

contraction “I will” & won’t for 1^st person instead

form “We will” of shall not / shan’t

I shall ---------> I’ll ----------> I will -------------------> I won’t

We shall ---------> We’ll ----------> We will -------------------> We won’t

Do not really

have contractions

Shall I ----------------> XXXX

Shall we ----------------> XXXX

Another possible explanation for the prevalence of shall over will (in question form but not the sentence form) is what I call the Contraction Hypothesis. This hypothesis proposes how will might have assumed the role of shall gradually through time. The hypothesis also accounts for why the replacement of shall by will only occurs in the sentence but not the question form. As illustrated in Diagram 4, we started with “I shall” and “we shall” at the beginning. However, people often talk (or maybe think) in contraction forms (I’ll and we’ll) instead. Since the two contraction forms sound more like “I will” and “we will”, through time, people might have grown to believe that will is the correct auxiliary verb to use for all three persons. This explains how will might have replaced shall in the sentence form. In contrast, since there are no contractions for “shall I” or “will I”, the question form did not experience a similar replacement of shall by will.

(3-c) Conclusions:

Comparing to the sentence form, the question form seems to give lower WSRs. In fact, for cases involving first person, the WSRs are actually smaller than one (i.e., shall is more commonly used than will). One possible reason for the difference (between sentence and question forms) is that in the question form, shall has additional meanings and applications. Therefore, there might be a higher frequency of shall (and thus a smaller WSR) in question form. An alternative explanation for the difference is the Contraction Hypothesis.

(3-d) Confirmation of Findings Utilizing Transitional Probability (t-score with span 0:1 or 1:0):

The “collocation” module in Cobuild was used in this part of the analysis. Shall and will were entered as node words and a span of 0:1 or 1:0 was used. T-scores (= transitional probabilities, in this case) were then calculated. The following are the results:

Table 5: Pronouns that Collocated Highly with Shall / Will and Their T-scores

Shall

Will

Span of 0:1

(Question form)

We (t-score: 22.279)

I (t-score: 18.65)

None of the pronouns makes the top 200 on the collocation list.

Span of 1:0

(Sentence form)

I (t-score: 30.55)

We (t-score: 28.81)

You (t-score: 6.32)

They (3.67)

He (1.68)

It (t-score: 82.71)

They (t-score: 69.18)

We (t-score: 66.89)

You (t-score: 66.89)

He (t-score: 62.08)

I (t-score: 41.28)

She (t-score: 28.30)

As demonstrated by the data in the first column, it is highly likely for shall to be preceded or followed by we and I versus other words (revealed by the high t-scores we and I have with shall). We and I are the two pronouns that top the collocation list for shall with span 0:1. This supports our finding that “shall I/we” is a fairly prevalent structure. The absence of the second and third person pronouns on this list confirms our other finding that shall is less commonly used with you/he/she/it/they in question form.

Looking at the collocation list for shall (with span 1:0), we see not only we and I, but also you, they, and he as being high collocates with shall. In other words, it is fairly likely that shall is preceded by these three pronouns. But, why do we still find such large WSRs for these pronouns then (70.8 for 2^nd person, 197.0 for 3^rd person singular, and 142.6 for 3^rd person plural)? The reason becomes apparent when one looks at the collocation listing for will (with span 1:0). It seems that the associations between will and you/they/he are even stronger (as indicated by the substantially higher t-scores) than these three pronouns’ collocation with shall. One last conclusion from this collocation list (will with span 1:0) is that we and I are highly collocated with will also (even more so than the collocation of will and I with shall). This affirms our finding that the traditional rules of usage for shall are fairly outdated (in the sentence form).

(4) Overall Distribution of Shall and Will Usage:

(4-a) Differences between spoken and written English:

We have previously looked at the overall WSRs for sentence and question forms (Table 3). In Tables 6 and 7, we are going to refine our focus and examine differences in WSRs across written (oznews, ukbooks, etc.), conversational (ukspok), and radio-spoken (bbc and npr) English.

Table 6: Variability of WSRs Across Written, Conversational, & Radio-Spoken English – Sentence Form

“Will / Shall” Ratio (= Instances of “Will” / Instances of “Shall”)

1^st Person Singular

“I”

1^st Person Plural

“We”

2^nd Person

Singular/Plural

“You”

3^rd Person Singular,

“He / She / It”

3^rd Person Plural

“They”

Overall

4.3

6.5

70.8

197.0

142.6

Written

4.5

7.3

68.5

169.0

146.2

UKSpoken

3.2

3.1

86.4

122.8

308.0

Radio

8.3

8.5

103.0

1161.0

75.3

As seen in Table 6, the WSRs for written English closely resemble the overall ratios. This is probably because the majority (74% based on number of words) of the Cobuild Corpus is written English. Another notable observation is that radio-spoken WSRs are generally higher than the other WSRs (except for third person plural). One possible explanation could be that the radio-spoken English provided in the corpus are from fairly high-register stations (BBC and NPR). Therefore, the lines are often scripted / prewritten (for example, the news). The scripts are probably composed in an even more formal style than general written English (which includes ephemerals and magazines) is. One point worth noticing is that although both conversational and radio-spoken English are considered “spoken”, they have very different WSRs. In other words, it is important for corpus linguists to distinguish between the two (rather than just examining them jointly under “spoken English”) when conducting research.

Table 7: Variability of WSRs Across Written, Conversational, & Radio Spoken English – Question Form

“Will / Shall” Ratio (= Instances of “Will” / Instances of “Shall”)

1^st Person Singular

“I”

1^st Person Plural

“We”

2^nd Person

Singular/Plural

“You”

3^rd Person Singular,

“He / She / It”

3^rd Person Plural

“They”

Overall

0.57

0.24

81.9

51.0

52.7

Written

1.2

0.58

51.1

38.0

82.4

UKSpoken

4.54

0.05

¥ [xi]

12.7

Radio

0.14

0.72

167.0

For second and third persons, the question-form WSRs, like the sentence-form WSRs, are significantly larger than one (see Table 7). However, the question form seems to have relatively lower written WSRs when compared to the sentence form. The lower WSRs could be, in part, attributable to the higher frequency of usage of “shall+he|she|it|they” in the literature or biblical context (most likely written).[xii] For instance:

Biblical:	If any will not work neither shall he* eat.*
Literature:	When shall he* come? Tell me, Othello.*

As for the first-person question form, both the singular and the plural have an overall WSR of less than one. For first person plural, shall is consistently preferred over will in written, conversational, and radio-spoken English (see Table 7). Nevertheless, for first person singular, the radio-spoken data seems to be the only factor driving the overall WSR to be less than one (since neither written nor conversational English has a WSR smaller than one). However, one should also take note that the even though the written WSR is not smaller than one, its value is substantially smaller in the question form than in the sentence form (1.2 and 4.5 respectively). In other words, the dominance of will over shall is weaker in the question form.

One last remark is that for both the sentence and question forms, there seems to be no consistent differences between conversational and written English (in terms of will and shall usage). Sometimes, written English has a larger WSR than conversational English. At other times, the opposite trend is observed.

(4-b) More on Differences between British and American English:

Table 8: “US / UK” Ratios for Different Persons / Number & Across Different Corpora

“US / UK” Ratio for Sentence Form

(“US / UK” Ratio for Question Form)

1^st Person Singular

1^st Person Plural

2^nd Person

Singular/Plural

3^rd Person Singular

3^rd Person Plural

Overall

2.7

(0.64)

1.1

(1.5)

0.031

(0.49)

0.063

(0.087)

0.22

(0)

Ephemeral

13.0

(0.63)

25.8

(¥)

1.3

(CBD)

0.87

(¥)

CBD

(CBD)

Books

2.5

(0.84)

1.54

(1.3)

0.21

(0.55)

0.12

(0)

0.13

(0)

Radio

0.96

(1.3)

2.83

(5.1)

(CBD)

(0)

In Section 2, we have examined the US-UK differences in using shall and will (for the sentence form). Looking at Table 8, we can draw similar inferences for the question form. For second and third persons, it seems that the USUKRs tend to be smaller than one (with a couple exceptions). As for first person plural, the sentence and question forms demonstrate a similar pattern to each other. The biggest difference between the sentence and question forms, therefore, lies in cases involving first person singular, when the question form gives similar, if not smaller, USUKRs than the sentence form.

(4-c) Table of Summary – Functions of Shall and Will for different persons and number:

The following table is compiled based on samples outputs from the Cobuild corpus (100 samples for each of the cells, except for cells with less than 100 outputs) and provides a brief summary of how shall and will are used and what functions the two words have in different cases.

Table 9: Functions of Shall and Will for Different Persons and Number – Sentence and Question Forms

Sentence Form

Question Form

___ + shall

___ + will

shall + ___

will + ___

1^st Person Singular

“I”

SF; EM

SF; PQ

1^st Person Plural

“We”

SF; EM

SF; PQ

2^nd Person Sing / Plur “You”

ORIT; EM *

SF; EM

SF; EM
(EM >> SF)

SF; PQ
(PQ >> SF)

3^rd Person Singular

“He / She / It”

ORIT; EM *

SF; EM

EM; ORIT *

SF; EM

(EM >> SF)

3^rd Person Plural

“They”

ORIT; EM *

SF; EM

EM; ORIT *

SF; EM

(EM >> SF)

SF = simple futurity

EM = a sense of emphasis

ORIT = orders, rules, instructions, threats, etc.

PQ = polite way of posing questions

A >> B = many more instances of A than B

* = this category consists of many biblical and/or literature examples

Final Conclusions:

A brief synopsis of our major findings is as follows:

(1) The traditional rules for shall and will usage are somewhat outdated (at least in the sentence form) – Will is used four to six times more often than shall in the first person. However, the preference of will over shall is significantly weaker in the first person than in the other two persons, suggesting that the traditional rules, although not as prevalent, may still be affecting people’s usage pattern on some level.

(2) There is no conclusive evidence as to whether the traditional rules are followed more closely in British than in America English. However, relatively speaking, the British does seem to have a slightly higher preference for (a) shall in cases of first person and (b) will in cases of second and third persons, comparing to Americans.

(3) In first-person question form, the dominance of will over shall is reversed. This observation might be explained by the “Variation in Meaning” Hypothesis. An example of the hypothesis is that unlike “will we”, which merely conveys futurity, “shall we” can also be used when asking questions or making suggestions politely. The additional functions of “shall we” could then have resulted in many more instances of “shall we” (even more than instances of “will we”). Another possible explanation for our observation is the Contraction Hypothesis (see Section 3 under Procedures, Results, and Discussion).

(4) The differences between conversational and written English (in terms of will and shall usage) show no consistent patterns.

(5) The patterns of will and shall usage in the radio-spoken corpora (bbc and npr) seem fairly different from those in the conversational corpora. Therefore, in conducting linguistics research with data from the Cobuild corpus, one might find it helpful to distinguish between the sub-categories of spoken English.

(6) Shall and will carry different meanings depending on (a) person (first, second, or third), (b) number (singular or plural), and (c) form (sentence or question) – see Table 9 for more details.

Before I end this paper, I would like to discuss a couple issues I came across during the course of my research. The first issue raised by this project is whether grammar rules dictate how people use language or the opposite is true. As seen in our findings, people often do not adhere to grammar rules rigidly and blindly. For example, as discussed earlier, will’s phonetic ease relative to shall may have contributed to its preference over shall in the first person (despite the prescription of traditional grammar rules). In other words, traditional rules (in this case, the rules for shall in the first person) can fall out of favor at times. I am not questioning the possible influence that rules have on usage pattern. In fact, as mentioned before, the smaller WSR for first person than for the other two persons may be due to the residual influence of old grammar rules. What I am pointing out though is that there are really no “eternal” rules in language usage. And the non-existence of a consistent set of grammar rules may imply that the competence argument (i.e., there is a set of pre-programmed rules of language in our brain) may not be valid.

Another interesting point that my analysis raises is the importance of corpus data in linguistics research. When I first told my colleagues about the topic of my research study, their immediate response is that shall is pretty much obsolete in the first person and that no one really uses shall for first person anymore. Their comments were based mostly on intuitive data (i.e., how many instances of shall used in first person that they could recall). My colleagues’ initial mode of reasoning, therefore, is somewhat similar to the traditional (Chomskyan) way of conducting linguistics research. The emphasis is on intuitive data and isolated sentences. However, as corpus data have revealed, our intuition (at least in this case) seems to be limited to shall in the sentence form. In the question form for first person, not only is shall not obsolete, it is actually more prevalent than will. In other words, if we have conducted our research based only on examples that we create artificially, we might have easily missed the distinctive patterns demonstrated by the question form. A more reliable method to conduct linguistics research would then be the use of corpus data.

Last but not least, future research can possibly look at should and would – the past tense (or at least what used to the past tense) forms of shall and will. This may be allow us to develop additional insights on why shall and will are used the way they are now.

References:

Jespersen, Otto (1964). Essentials of English Grammar. University, AL: University of Alabama Press.

Quirk, R. and S. Greenbaum (1973). A Concise Grammar of Contemporary English. New York: Harcourt Brace Jovanovich.

The American Heritage Dictionary of the English Language (1996). 3^rd Edition. Houghton Mifflin Company. http://www.bartleby.com/64/C001/056.html