You spoke recently about the fact that our country habitually fails a third of its pupils in Maths and English after more than 10 years of compulsory schooling.
I don't know anyone who does not agree that this is a national disgrace that has been going on for way too long.
You agree that, compounding the shame and depair of being branded as failures, we then force those 300,000+ young people each year onto a 'hamster wheel' of unnecessary resits where only one-fifth (20%) manage to get to the required standard.
It does not take a genius to conclude that we have got something horribly wrong. I see three possible causes:
(a) schools are terrible?
(b) one third of young people can never achieve an acceptable national standard in numeracy and literacy?
(c) our assessment model for literacy and numeracy is not fit for purpose?
I've been in education for 14 years (after a long career in business), and I can say with confidence that our schools are generally pretty good. So that rules out (a).
Literacy and numeracy are fundamental parts our humanity. Reading, writing and applied numeracy underpin our ability to survive and thrive in the world. If we have come to perceive literacy and numeracy as "academic" pursuits, suited only to two-thirds of the population, then we must be presenting and measuring them in the wrong way. I simply do not accept that we cannot get 90+ percent of our children to a good national standard of literacy and numeracy in which they can take pride.
Many schools (including schools where I am an advisor/trustee) set 90% goal as a target, and some achieve it. But obviously not all can - the system prevents it.
So, if (b) were the cause, based on the government's assumption that our national standards for numeracy and literacy are "too academic" for one third of young people, that would be catastrophically damaging and unecessarily cruel. So I'm going to rule out (b) on the grounds that we are not so inhumane.
The latest government response to the so-called 'forgotten third' is to suggest that those who are unlikely to ever get a grade 4 in English and Maths should be given an technical-vocational alternative from age 14. I'm certainly not against early identification of technical/vocational pathways for a sizeable minority of young people who really want to follow that route for the right reasons, and especially if both carrot and stick are used to get employers to properly commit to apprenticeships. But that is really solving a different problem, in many cases for children who would actually get their Grade 4+ passes anyway. We cannot ignore that technical-vocational pathways still need a good standard of numeracy and literacy.
I say all of this as an English graduate (and late-career English teacher) and as a chartered accountant who has been a CFO of several multinational companies.
And all of this leads me to conclude that (c) must be the cause of the problem.
Chaired by Roy Blatchford CBE and published by ASCL in 2019 The Forgotten Third drew attention to this national disgrace.
A truly groundbreaking report, it influenced many working in and outside the education sector, myself included. It made some very good recommendations but few (if any) have been acted upon. It is still very regularly cited and is more relevant now than ever.
For me, its four most powerful recommendations are:
1) put oracy at the heart of the curriculum,
2) stage not age - assess children when they are ready, not when it suits the system.
3) criterion-referenced assessment. Move away from "comparable outcomes" towards assessments where students are judged against specific standards of learning and competency. This makes 90+ percent of attainment more achievable. An analogy with the driving licence is a good one. Driving is a highly evolved and complex set of cognitive competencies, yet over 90% of those who start the process pass it eventually. Functional necessity is a powerful stimulus.
4) everyday applied skills. This continues to be a broad and contested area. For me, applied numeracy* would include spreadsheet skills (largely ignored in the current system) for things like personal budgeting and and project planning. Literacy/oracy would include effective social and professional interaction for relationship-building, problem-solving and critical thinking. Portolio submissions and context-driven project qualifications would form part of the equation.
*though not mentioned in the report, computational literacy is also a strong candidate for category 4. Sometimes described as 'maths that acknowledges the ubiquity of the computer', it is the ability to use computational thinking, tools and concepts to solve problems, analyse data and understand how digital systems work.
We learn to listen and speak well before we can read and write. Yet we don't formally recognise this essential part of human development in our education system.
I have written extensively on oracy, including in other parts of this website, so I won't repeat it all here. I'll just say that people often get confused between two quite different things:
oracy skills
oral assessment as an effective method to evaluate knowledge and understanding across all subjects. It is especially effective in the age of AI, and especially for many SEND pupils. To be clear, oral assessment is not just about assessing oracy skills.
Absent from the interim report, oracy thankfully made it into the final report of the government-commissioned Curriculum and Assessment Review (2025). But its presence is weak with no concrete measures being laid out for formal assessment and accreditation.
Oracy has always been underrepresented in our national currriculum and assessment, but with the advent of generative AI, the demand to give oracy parity with literacy is growing very rapidly.
This is because:
traditional written assignments and assessments are much less reliable following the widespread adoption of generative AI. As the line between human and machine authorship becomes impossible to determine, listening and speaking may become the best evidence of our human authenticity, empathy and capacity to connect with each other.
AI is already making a huge contribution to pupils with special learning needs, through tools such as voice-typing and text-editing.
accrediting oracy equally with literacy is arguably the most inclusive policy we can implement because it recognises that many are excluded by their struggles (for various reasons) to develop literacy at the same pace as others in the same age group. Oracy skills evolve independently from literacy skills - recognising them formally can provide much needed confidence to many children.
some would argue we have already entered a post-literate age, where appetites for reading and writing long-form text are in decline. Effective speaking (including using voice-instruction and agentic AI delegation for writing) will erode the historic premium put on unassisted long-form writing skills.
The obstacles to putting oracy at the heart of our curriculum and assessment systems are both cultural and structural. AI gives us the opportunity to address both, but only if we build oracy in a low stakes way from a young age.
The prevailing orthodoxy of high-stakes exams at age 16 as the only "gold standard" has seemed obsolete for some time.
For those children whose abilities and family (or tutor-enhanced) support mechanisms are suited to the system, it feels marvellous to achieve eight or more good passes at GCSE. But the playing field is not level. As we all realise when the immediate euphoria of exam success wears off, the positive influence of GCSE results has a short shelf-life.
And for the third that fail, the influence can be catastrophically long-term.
The Forgotten Third introduced the concept of the Passport in English. Like the driving licence, it would be designed such that:
a) students take it when they are ready
b) it would have a practical (demonstrative) basis of evaluation (oral assessment)
c) over 90% of people would achieve the standard
d) it would make students proud, and confident that it would help them thrive in the world.
In no way a 'dumbing down' of numeracy and literacy, but a more obvious manifestation of their practical utility.
English language, literature and mathematics are beautiful and powerful intellectual pursuits. They should remain central to our educational system. But we should not conflate their majesty with what is most needed as a minimal national standard of literacy, oracy and numeracy for all citizens to sustain their dignity.
An annual combined oracy-literacy test from age 6, ensuring all children progress having met a minimum standard. This way the 95% of children might achieve their Passport in English to survive and thrive with dignity in the world.
It is rooted in the assumption that young children learn to listen and speak before they can read and write. At primary school, assessment would be designed as low stakes to normalize the activity of oral assessment and avoid anxiety or stress (for pupils and parents).
Using the capabilities of generative AI, children would interact with a Socratic AI assessor, appropriately supervised and moderated by qualified human educators.
The assessment would follow a four-stage process:
Auditory engagement
Socratic dialogue
Iterative writing (mirroring the 'planning and drafting' skills often neglected in current assessments)
Vocal performance
To begin with pupils would listen to a short text (no more than 200 words).
Through oral interaction with the Socratic AI chatbot, they would then be prompted to explore and express their thoughts orally on the text, assessing their comprehension, vocabulary, ability to engage in dialogue and their imagination.
After this oral phase is completed, they would then be asked to write about what they have learned and felt (no more than 200 words).
All children (or perhaps just children less comfortable with writing) would be given the option of voice-typing. In either case, the child would be given the option to correct their typed text, using appropriate edit technology for suggested changes. Though inverted, this builds on the old pedagogical practice of dictation as an effective method of language acquisition. Indeed, dictation could easily be reintroduced into schools as research suggests that its mid-20th century dismissal as a "passive drill" was misplaced.
In the final stage of the assessment pupils would read out loud the text they have written. This would be assessed across a number of criteria including clarity of expression and diction.
This model faces significant cultural and political obstacles. Politically, the shift would require a major change in accountability measures and substantial investment in technology and teacher training, which may raise concerns regarding system workability.
However, this approach offers transformative advantages through the formal accreditation of oracy.
At the very least, it is surely worthy of discussion?
*I've limited myself to literacy and oracy as I don't feel sufficiently pedagogically-qualified to propose a solution for numeracy. Though I'd be happy to do so.