One finding, eight sentences
The same finding — that emissions have fallen — written eight ways. They are not degrees of the same thing. Each one is uncertain about something different.
Eight sentences. Match each to the part of the claim it puts in doubt.
Not a dial — a taxonomy
Six things you can be unsure about
Hedging is usually explained as a way of sounding less certain. That is true and it is unhelpful, because it suggests there is one hedging move and a dial to set it on. There is not. English has at least six different things you can be unsure about, and a different instrument for each of them.
The fact: it seems that… · the evidence: …appear to have… · the probability: probably · may · is likely to · the frequency: tend to · in general · the scope: in most cases · in industrialised economies · at this scale · the source: according to one study · the 2021 review found. And a seventh worth counting separately: the magnitude — roughly · in the region of · at least.
Which is why the advice everybody is given — add maybe, write I think — produces bad academic prose. Both of those attach to the fact and to nothing else. They put doubt on the one thing you may well be sure of, and leave the scope, the source and the size of the effect claimed at full strength, which is usually exactly the wrong way round.
The second half of this is about how a reader takes it, and it is the part that surprises people. An unhedged claim in a scientific or policy register does not read as confident. It reads as unaware — because a reader who knows the field can see the exceptions you have not mentioned, and concludes that you cannot see them. Hedging is the clearest single marker of expertise in academic English, and it is not a marker of timidity at all.
That is the specific problem for Portuguese speakers, and it is not simply that you under-hedge. Talvez and eu acho que cover the whole space, so the choice of instrument never has to be made — and in conversational Portuguese those moves do read as hesitancy, which makes dropping them feel like an improvement. In English academic prose it is the reverse.
So today counts two things: how many absolute claims you make in fifteen minutes, and how many different hedging instruments you use. Most people at this level score high on the first and one — occasionally two — on the second.
Four questions.
1. What is wrong with “add maybe” as advice?
2. How does an unhedged claim read to somebody who knows the field?
3. What is tend to doing?
4. Why do Portuguese speakers under-hedge in English?
The six slots, and six sentences that miss them
The six slots, one example each.
Fact — it seems that / it would appear that
Evidence — X appears to / X seems to
Probability — may · might · is likely to · probably
Frequency — tend to · generally · in most cases
Scope — in industrialised economies · at this scale · among adults
Source — according to · one study found · the 2021 review reported
And magnitude, which behaves like a seventh: roughly · in the region of · at least · no more than.
Six sentences. Four claim more than the evidence supports; two hedge so much that they claim nothing at all. The note tells you what is actually known.
1. Renewable energy is always cheaper than gas. — true in most markets since about 2020, not in all
2. Everyone knows that recycling does not work.
3. I think maybe possibly the data could perhaps be wrong.
4. Scientists proved that the ice sheet will collapse. — one modelling study, one scenario
5. This study shows that electric cars are bad for the environment. — one study, one country, one manufacturing route
6. It seems that it could possibly appear to be the case that emissions have perhaps fallen.
One study, several claims
Eight gaps. Every option is grammatical. Choose the one that claims exactly as much as the writer can support.
A study published last year that air quality in the city centre has improved since the charging zone was introduced. The improvement to be concentrated in the morning peak, which is what the model predicted, and it holds of the eleven monitoring sites. Nitrogen dioxide fell by twenty per cent at the two busiest junctions.The paper is careful about what it does not show. Traffic to fall in the first year of any charging scheme and then to recover, so a single year be enough to establish a trend. The authors note that particulate levels, which come largely from brakes and tyres rather than from exhausts, unchanged. And the same design has been trialled in three other cities, none of them has published results yet.
One word per gap. Eight claims, each hedged in a different slot.
The evidence to suggest that urban air quality has improved since 2015.
Emissions to have fallen in most industrialised economies.
Cities of this size to have worse air quality than the national average.
The improvement is to be temporary, on the evidence so far.
This pattern holds in European capitals, though not in all of them.
The reduction was four per cent, depending on how it is measured.
One study that the effect may be seasonal.
The published figure well be an underestimate.
The biggest environmental challenge
Task card · Speaking · Pairs · 15 minutes
Three minutes each on the environmental question you actually care about — air, water, waste, transport, food, energy, land. Your partner is not listening for opinions.
Two counts. Absolute claims made, and different hedging instruments used.
- 4 minutes alone: your question, why it matters, and three claims you would make about it. Write the claims down.
- 3 minutes each, speaking: make the case. Do not edit yourself — today’s data is what comes out unprompted.
- Partner, column one: tally every absolute claim — anything with always, never, everyone, proves, or a bare present simple stating a contested fact.
- Partner, column two: list the different hedging instruments used, not the number of hedges. Write the two figures at the top of the page.
Expect column two to be short. One instrument is normal at this level, two is good, and the instrument is almost always I think. That is not a personal failing — it is what happens when one word in your first language covers six slots in your second.
Two columns
Three questions.
1. How many different hedging instruments did you use?
2. Which slots did you leave completely unhedged?
3. Did the absolute claims feel confident as you said them?
Which part are you unsure about?
Two questions. Twenty seconds.
1. What does a hedge do?
2. How does an unhedged claim read to an expert reader?
Next. Four texts about the same research — a feature, a press release, the abstract itself, and a comment thread — and a count of how much each one claims.