I was going to post something on the messy and increasingly worrying business of Brazilian politics, but instead I find myself seeking solace in a more abstract comfort zone of statistical linguistics and discourse analysis

According to Google N-gram, the pattern of frequency of the nine main Modern English modal verbs, can, could, may, might, must, shall, should, will and would, in texts made available online by GoogleBooks, has changed significantly over time.

With the exception of can and could, the frequency of all modal verbs in this corpus declined between 1800 and 2000. At the outset of the 19^th century, these nine modal verbs accounted for 1.09% of all lexemes in the corpus. By 1900 this had dropped to 1.02% and by the year 2000 to 0.87%

The most striking change is the rise of can. In 1800 and 1900, can ranked fourth behind will, may and would with a frequency of 0.13%. By 2000, however, it had risen to a clear first place with 0.20%, 0.06 percentage points ahead of will and would in equal second place. Most of this change occurred in the second half of the 20^th century.

The frequency of could remained unchanged over the 200-year period surveyed, although, given the overall decline in the use of modal verbs, this translates into a rise from 8^th place out of 9 in 1800 to 6^th in 1900 to 5^th in 2000.

The modal verb that saw the sharpest decline was shall, falling from a frequency of 0.09% (7^th place) in 1800 to 0.7% (8^th place) in 1900 and 0.2% (last place) in 2000. May also declined, albeit at a slower pace, from 0.16% (2^nd place) in 1800 to 0.15% (3^rd place) in 1900 and 0.11% (4^th place) in 2000.

These bald statistics merit some discussion and interpretation.

Given, as always, the proviso that the kinds of texts included in the Google corpus may differ according to historical period (e.g. more fiction in earlier years, more non-fiction recently), thus biasing the results, the changes revealed would appear to be related to shifts in meaning and fluctuating trends over time and, more controversially, to ongoing social, cultural and technological developments.

The large increase in the use of can may be explained in part by the tendency, especially over the past 50 years, to fail to distinguish between can (=is physically able) and may (=is permitted) and to use can for both purposes. In more old-fashioned language “Can I open the window?” means “Am I physically capable of opening the window?” rather than “Do you mind if I/am I allowed to open the window?” While this distinction may persist to this day in the politest of spoken discourse, it is almost gone from written texts. “May I open the window?” sounds OK, if a little overweening, “Immigrants may apply for a permanent visa within 90 days” has a somewhat fussy, fusty and off-putting tone, compared to the same sentence with may replaced by can.

The greater frequency of can could also be explained by the rise, especially since the 1950s, spurred by the advertising and self-help industries, of a mentality of “can do”, “ yes we/you can”, often with a previously unnatural sentence stress on the modal verb. This reflects both a rise in personal liberties, but also, more negatively, an increasing insistence that people draw on their supposed inner resources for personal betterment, rather than rely on others: “You can lose weight…” “You can quit smoking,” “You can be successful…”

In an age of increasing inequality and powerlessness in the face of addictions and the multinational corporations that fuel them, the growing use of this little word conveniently shifts the lion’s share of the blame onto the victim, without resorting to the harshness of modal verbs of obligation–“must, should, shall,”—all of which have declined significantly in frequency of use between the early 19^th and the early 21^st centuries.

Google’s frequency analysis does not allow us to distinguish between the modal verb ‘can’ and ‘can’ as a noun meaning a container for processed food and beverages. Both of these could reasonably have been expected to have increased in frequency in the period after the 1950s—and, to some extent, for the same reasons.

We can measure the potential bias produced by the emergence of canned food and drinks by comparing the frequency of can (modal verb + substantive” with that of near synonyms that have no grammatical functions.

“Tin” for example is synonymous with ‘can,’ albeit rarer in American English, and also has another substantive meaning, referring to the metallic element whose chemical symbol is Sn, which would need to be subtracted from the total. Rounding numbers down to two decimal places, as I have done with the modal verbs analyzed here, ‘tin’ never rises above 0.00% throughout the 200-year period. It reached a peak of 0.002% in the late 1940s, but has since declined by 50%, probably because of the rise of “can”. This would suggest that the contribution of “can” as a substantive to the overall frequency of the word is minimal.

A comparison with “box” shows an equally gaping disparity. “Box” has a frequency of about 0.005% up to the late 1980s, when it surges to 0.01% and, as with ‘tin,’ a small measure of this must be accounted for through use of ‘box’ as a verb in connection with the sport. However, even if “can” as a substantive enjoyed an equivalent increase over the same period, this would still only account for a minimal fraction of its growing frequency.

Even more convincingly, the word “computer”, which one would expect to rise in frequency dramatically in recent years, especially in a Google-gathered and hence biased corpus, shows exactly the same pattern as “box,” increasing from a low level of less than or equal to 0.005% up to the 1960s to a relatively meagre 0.01% at the turn of the millennium compared with modal verbs. It can therefore be concluded that the influence of canned food and fizzy drinks on the rising use of can as a modal verb is minimal and insignificant.

The reverse however may not be true. Google’s corpus does not allow us to compare the global frequency of use of the words “can”, “tin,” or the more exotic Australian “tube,” to refer to a standardized cylindrical container for processed food or drink. However, there can be little doubt that “can” has won the world-wide competition for the more popular term. This may be largely due to the sheer weight of numbers in favor of US or US-influenced “can” compared to parochial UK “tin” and Australian “tube”, but surely also has something to do with the apt convergence of the term with a growingly popular modal verb that signifies personal freedom and ease of use.

American English peculiarities do not, therefore, spread simply by dint of the economic, military and demographic weight of the US. Clunky US “faucet”, by contrast, is clearly parochial compared to the nicer and simpler “tap,” which has now taken on a wide range of different meanings, functions and grammatical forms around the world.

The global psychosocial shift towards language that implies ease of achievement and ease of use may also in part explain the rapid decline in the frequency of shall (with its overtones of external authority) compared to will (with its overtones of personal volition) in the Google corpus.

“You shall lose weight” (a phrase that sounds odd, if not outright wrong, in modern discourse) implies that the speaker has some magical power or authority to make this happen, while “You will lose weight…” implies that this is some natural process; hedges the statement—and hence dilutes the responsibility—by couching it in the form of a prediction regarding an always uncertain future; and, more subtly, suggests that the patient object of treatment’s own will is somehow intimately and ultimately entwined as subject and agent in a radically alien imposed healing process.

Apart from this, shall may have been over-represented in earlier texts because of adherence to the erroneous prescriptive rule that it is the correct modal verb to use to express the future in the first person.

In many empirical examples, this prescriptive use (albeit erroneous) does in fact overlap with the authoritative overtones of shall, especially when used by a speaker in a position of power, as most people who knew how to use it ‘properly’ at the time would have been.

“We shall fight them on the beaches!” conveys a much more powerful sense of duty and resolve than the somewhat lily-livered-sounding ‘will’—‘we want to, but…’. ‘maybe we will…’. To cite another Churchill anecdote concerning modal verbs: in the aftermath of the Japanese attack on Pearl Harbor, the British Prime Minister’s first words to Roosevelt on the trans-Atlantic telephone were “What shall we do?” Here shall has much more exhortative power than a helpless-sounding deer-in-the-headlights “What are we going to do?” Will here, interestingly, would simply be wrong.

While there may be some value in the decline of arbitrary authority and grammatical prescription in the modern world, the reduced frequency of shall and the irresistible rise of can also bear witness to a certain failure of resolve, combined with a growing tolerance of whimsical self-interest and comfort at the expense of collective duties and needs and an unhealthy (often hopelessly idealistic) emphasis on what underprivileged individuals can or should do. Politicians nowadays are more likely to tell us what they will do (i.e. what they want), what is going to happen (a shrug of inevitability) or what they or we can do (bombast in the first person; victim-blaming dolled up as empowerment in the second).

We all can and should do better than this.