Speculation across various scientific topics

The searches

An earlier draft of this post ran the searches through a set of functions that I had written around the rscopus package. Those functions have since become the scopusflow package, and the searches now live in a script in the website’s repository that a GitHub Actions workflow runs on request, with the Scopus key held as a repository secret. The workflow commits the counts to the post’s directory, so the figures below can be rebuilt from the committed data without a key, and the searches can be refreshed by anyone with a key by dispatching the workflow. The nine queries are given below.

library(scopusflow)

years <- 1980:as.integer(format(Sys.Date(), '%Y'))

topics <- c(
  'language evolution'              = '"language evolution" OR "evolution of language"',
  'language comprehension'          = '"language comprehension"',
  'language disorders'              = '"language disorders" OR "language disorder"',
  'linguistic relativity'           = '"linguistic relativity"',
  'language teaching'               = '"language teaching"',
  'bilingual advantage'             = '"bilingual advantage" OR "bilingual advantages"',
  'sensorimotor simulation'         = '"sensorimotor simulation"',
  'artificial general intelligence' = '"artificial general intelligence"',
  'hadron collider'                 = '"hadron collider"'
)

comparisons <- lapply(names(topics), function(label) {
  cmp <- scopus_compare_topics(
    reference_query  = topics[[label]],
    comparison_terms = 'speculat*',
    years            = years,
    field            = 'TITLE-ABS-KEY'
  )
  cmp$topic <- label
  cmp
})

The function counts, for every year, the records matching the reference query and the records matching both the reference query and the comparison term, so the nine comparisons cost 846 count requests. Each row of the result holds a year, the count for the reference literature, the count for the comparison and the comparison as a percentage of the reference.

library(dplyr)
library(ggplot2)

comparisons <- readRDS('searches/speculation_comparisons.rds')
retrieved <- attr(comparisons, 'retrieved')
retrieved

#> [1] "2026-09-02 06:36 UTC"

speculation <- bind_rows(comparisons) |>
  filter(query_type == 'comparison') |>
  transmute(topic, year, reference_n, n,
            percentage = comparison_percentage)
head(speculation)

#> # A tibble: 6 × 5
#>   topic               year reference_n     n percentage
#>   <chr>              <int>       <dbl> <dbl>      <dbl>
#> 1 language evolution  1980           5     0          0
#> 2 language evolution  1981           4     0          0
#> 3 language evolution  1982           4     0          0
#> 4 language evolution  1983           1     0          0
#> 5 language evolution  1984           5     0          0
#> 6 language evolution  1985           6     0          0

The searches were run on 2 September 2026, so the final year is incomplete and its share rests on fewer records than the years before it.

Sizes of the literatures

The nine literatures differ in size by orders of magnitude, and the size matters for reading the percentages. A share computed on a few dozen records moves by several points when one or two abstracts mention speculation, whereas a share computed on thousands of records barely moves at all.

totals <- speculation |>
  group_by(topic) |>
  summarise(records = sum(reference_n), with_speculation = sum(n),
            share = 100 * with_speculation / records,
            first_year_with_records = min(year[reference_n > 0])) |>
  arrange(desc(share))
totals

#> # A tibble: 9 × 5
#>   topic                           records with_speculation share first_year_with_records
#>   <chr>                             <dbl>            <dbl> <dbl>                   <int>
#> 1 artificial general intelligence    1913               49 2.56                     2006
#> 2 language evolution                 3376               48 1.42                     1980
#> 3 linguistic relativity               739                9 1.22                     1980
#> 4 bilingual advantage                 637                4 0.628                    1993
#> 5 language comprehension             6543               24 0.367                    1980
#> 6 language disorders                14255               36 0.253                    1980
#> 7 language teaching                 22957               42 0.183                    1980
#> 8 hadron collider                   18457               20 0.108                    1980
#> 9 sensorimotor simulation             132                0 0                        2003

Across the whole period, the share of records that mention speculation is highest in the literature on artificial general intelligence (2.6%) and lowest in sensorimotor simulation (0.0%), with the language literatures spread between those two.

# Nine literatures against an eight-colour palette, so the Okabe-Ito set is
# extended with one grey. Nine lines on one panel are hard to tell apart by
# colour alone, which is why the figure that follows gives each its own panel.
nine_colours <- c('#E69F00', '#56B4E9', '#009E73', '#F0E442', '#0072B2',
                  '#D55E00', '#CC79A7', '#999999', '#000000')
ggplot(filter(speculation, reference_n > 0),
       aes(year, reference_n, colour = topic)) +
  geom_line(linewidth = 0.8) +
  scale_y_log10(labels = scales::label_comma()) +
  scale_colour_manual(values = nine_colours) +
  labs(x = NULL, y = 'Records per year (log scale)', colour = NULL) +
  theme_minimal(base_size = 12) +
  theme(legend.position = 'right')

The share of records that mention speculation

Each panel below shows one literature. The line is the yearly percentage of records that also match speculat*, the shaded band is a stability range computed by the Wilson method from the two yearly counts, which conveys how much the share would move with a handful of records, and the panels are ordered from the most to the least speculative literature over the whole period. The axis is cut at 15%, which clips the bands in the early years of the small literatures, where a single abstract would move the share by tens of points. The highest share in the data, the 33% that 1 of the 3 records on bilingual advantage in 2006 represent, sits above that cut and is clipped with them. Years in which a literature has no records break the line.

wilson <- function(k, n, z = 1.96) {
  p <- k / n
  centre <- (p + z^2 / (2 * n)) / (1 + z^2 / n)
  half <- z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2)) / (1 + z^2 / n)
  data.frame(lower = 100 * pmax(0, centre - half), upper = 100 * pmin(1, centre + half))
}

# Years with no records keep their row but carry NA, so that ggplot2 breaks the
# line there instead of drawing a segment across a gap in the literature.
panel_data <- speculation |>
  mutate(topic = factor(topic, levels = totals$topic)) |>
  bind_cols(wilson(speculation$n, speculation$reference_n)) |>
  mutate(across(c(percentage, lower, upper), ~ ifelse(reference_n > 0, .x, NA_real_)))

panel_labels <- totals |>
  mutate(label = sprintf('%s\n%s records, %.1f%% with speculat*',
                         topic, format(records, big.mark = ','), share))
labeller_topic <- setNames(panel_labels$label, panel_labels$topic)

ggplot(panel_data, aes(year, percentage)) +
  geom_ribbon(aes(ymin = lower, ymax = upper), fill = '#0072B2', alpha = 0.15) +
  geom_line(colour = '#0072B2', linewidth = 0.7) +
  geom_point(colour = '#0072B2', size = 0.9) +
  facet_wrap(~ topic, ncol = 3, labeller = as_labeller(labeller_topic)) +
  scale_y_continuous(labels = scales::label_percent(scale = 1)) +
  coord_cartesian(ylim = c(0, 15)) +
  labs(x = NULL, y = 'Records mentioning speculation (%)') +
  theme_minimal(base_size = 12) +
  theme(strip.text = element_text(size = 10, hjust = 0),
        panel.spacing = unit(1.2, 'lines'))

recent <- speculation |>
  filter(year >= 2016, year <= max(year) - 1) |>
  group_by(topic) |>
  summarise(records = sum(reference_n),
            share = 100 * sum(n) / sum(reference_n)) |>
  arrange(desc(share))
recent

#> # A tibble: 9 × 3
#>   topic                           records  share
#>   <chr>                             <dbl>  <dbl>
#> 1 artificial general intelligence    1377 2.54  
#> 2 linguistic relativity               374 1.60  
#> 3 language evolution                 1734 1.56  
#> 4 bilingual advantage                 445 0.449 
#> 5 language comprehension             3298 0.334 
#> 6 language disorders                 6511 0.230 
#> 7 language teaching                 15193 0.178 
#> 8 hadron collider                    7682 0.0781
#> 9 sensorimotor simulation              88 0

Restricting the comparison to the last ten complete years, 88 to 15,193 records per literature, gives the ranking above. The literature on artificial general intelligence leads with 2.5% of its records mentioning speculation, and that on sensorimotor simulation closes the list with 0.0%.

What the counts can and cannot show

The measure is crude in ways that the figures make visible. A form of ‘speculate’ in an abstract is as often a mark of candour as of conjecture, since the authors who write ‘we speculate that’ are labelling a claim that others would present without a label, and a literature with a strong convention of hedged abstracts will score higher for that reason alone. The wildcard also catches ‘speculative’ in senses that have nothing to do with scientific conjecture, such as speculative execution in computing or speculative investment in economics, and both senses can occur in the literatures on artificial general intelligence and on the hadron collider, whose funding is discussed in economic terms. The queries define each literature by a phrase, so records that belong to a field without using its name are missed, and the literatures on language teaching and language disorders are far larger than their phrases suggest. Percentages for the early years, when a literature had a handful of records, are shown with their intervals so that a single abstract is not read as a trend.

The smallest literature, sensorimotor simulation, contains no record that mentions speculation at all, which with 132 records over the whole period says more about the size of the literature than about its habits.

With those limits in mind, the pattern across the language sciences is consistent with the impression that started the exercise. Literatures that theorise about unobservable or contested processes carry more explicit speculation than literatures organised around a clinical population, a classroom or an instrument, and the difference is visible in abstracts before any reading of the articles themselves. Whether the speculation is a symptom of immaturity or a healthy habit of labelling conjecture is a question that counts cannot answer. Anyone who wants to add a topic, or to see how the shares have moved, can dispatch the workflow again from the repository.