speech markers

• Cambridge Dictionary +Plus

Discourse markers ( so, right, okay )

Discourse markers are words or phrases like anyway, right, okay, as I say, to begin with . We use them to connect, organise and manage what we say or write or to express attitude:

[friends are talking]

A: So , I’ve decided I’m going to go to the bank and ask for a car loan . B: That sounds like a good idea . C: Well , you need a car . B: Right . A: Anyway , I was wondering if either of you would teach me how to drive .

The discourse markers in this extract have a number of uses: so marks the beginning of a new part of the conversation.

well marks a change in the focus (from getting a car loan to needing a car).

right marks a response (B is agreeing with C).

anyway marks a shift in topic (from buying a new car to having driving lessons).

We use different discourse markers in speaking and writing. In speaking, the following discourse markers are very common:

In writing, the following discourse markers are common:

Discourse markers do not always have meanings that you will find in your dictionary. However, they do have certain functions, and some discourse markers, such as well , can have a number of functions.

Actual and actually

Discourse markers that organise what we say

Some discourse markers are used to start and to end conversations. Some are used to start new topics or to change topics.

Starting a conversation or talk

A: Right , let’s get started. We need to get the suitcases into the car . B: Okay . I’ll do that. Katie, will you help me?

[at the start of a radio interview]

Now , we have with us in the studio today someone you will all know from television. John Rice, welcome to the show.

Ending a conversation

[A mother (A) and daughter (B) on the telephone]

A: So we’ll see you Sunday, Liz . B: Right , okay Mum .

A: Okay , see you then, love . B: Bye, Mum. Thanks for calling . A: Bye, Liz .

[At the end of a meeting]

A: Anyway , is that it? Has anyone got any questions? B: No. I think we’re done . A: Right , fine, thanks everyone for coming. We’ll circulate the documents tomorrow and make some follow-up calls about the project .

Changing or managing a topic

A: We went to town to buy wallpaper to match the carpet . B: Did you try Keanes? They have a sale .

A: We looked there, but Jim said he thought it was too expensive and he didn’t like any of their designs . B: What does he like? A: He likes geometric shapes. He hates flowers . Anyway , we eventually found some that we both liked and when we went to pay for it, we realised that neither of us had brought any money . ( Anyway marks a return to the main topic of buying wallpaper.)

Ordering what we say

We also use discourse markers to order or sequence what we say. Some of the common words and phrases which we use for this are:

A: I think Sheila might be having some financial problems at the moment . B: I don’t think so, Caroline . For a start , she has all the money that her aunt gave her. What’s more, she has a good job and she seems to have a good lifestyle .

Firstly , we are going to look at how to write an essay. Secondly we are going to look at what makes a good essay and what makes a bad one. Lastly , we’re going to do some writing activities.

We can use the letters of the alphabet ( a, b and c ), to list reasons or arguments for something:

There are two reasons why I think it’s a bad idea, a because it’ll cost too much money, and b because it’ll take such a long time.

Numbers: first , second , third

Discourse markers that monitor what we say

As we talk, we monitor (or listen to) what we are saying and how our listener is responding to what they hear. We often rephrase or change what we say depending on how our listener is responding. We use words and phrases such as well, I mean, in other words, the thing is, you know, you know what I mean, you see, what I mean is .

Saying something in another way

Sometimes, as we talk, we add phrases to show our listener that we are going to rephrase, repeat or change what we are saying. These discourse markers help to make what we say clearer for the listener:

I just had to leave early. What I mean is I hated the show. It just wasn’t funny.

You exercise regularly, you have a good diet and you don’t have too much stress. In other words , I think you have nothing to worry about. Your health seems very good.

I think I’ve found a house I’d like to buy. Well it’s an apartment actually . It’s ideal for me.

Shared knowledge

When we talk, we think about how much knowledge we share with our listener. We often mark what we think is old, shared or expected knowledge with you know and we mark new knowledge that we see as not shared with the listener with phrases like see, you see, the thing is :

You know , hiring a car was a great idea. (The speaker and the listener know about hiring the car.)

A: Why don’t you come and stay with me when you’re in Lisbon? B: It’d be difficult. I have to be back in Dublin by Friday . You see , my sister is getting married on Saturday so I won’t have time to visit . (B assumes that A doesn’t know about her sister’s wedding. This is new information)

Discourse markers as responses

As we listen to someone speaking, we usually show our response to what we hear either by gesture (head nod) or by a short response ( Mm, yeah, really, that’s a shame ). This shows that we are listening to and interested in what is being said. We call these short responses ‘response tokens’.

Common response tokens include:

We use response tokens for a number of functions:

A: So he opened the door . B: Yeah . A: And he went in very quietly without waking her . B: Right . A: He opened her bag and…

To show surprise

A: We’ve decided to go to Africa for a month next year . B: Oh really!

To show sympathy

A: He can’t play soccer for at least six months. He’s broken his leg . B: That’s terrible .

Adverbs as discourse markers ( anyway , finally )

Discourse markers showing attitude

Some expressions are used to mark attitude or point of view in speaking or writing.

Common expressions of attitude are:

If you ask me , Neil is making a big mistake leaving his job to go travelling with his friends.

We will obviously have to pay for the damage done to the window.

The whole problem has been caused, I think , by having too many cars on the road at busy times.

Sadly , Hilda has decided not to come with us.

Discourse markers: sounding less direct

We are careful when we speak not to sound too direct or forceful. We use words and phrases such as like, maybe, sort of to soften what we say (hedges).

We often use these words and expressions as hedges:

* sort of is more common in British English ; kind of is more common in American English .

Can I just ask you a question?

We can probably add some more water to the sauce.

Is this perhaps one of your first times driving a car?

Hedges ( just )

Discourse markers: um and erm

We can use um to introduce a new topic carefully:

Um , could I ask you a personal question?

Um , there’s something else we need to talk about.

We can use erm when we pause before saying something, especially when we are not sure about what to say:

He’s… erm he’s not very pleased with your work, I’m afraid.

Her last book was called… erm what was it? I can’t remember the name.

Discourse markers: interjections ( Oh! Gosh! )

An interjection is a single-word exclamation such as hooray, oops, ouch which shows a positive or negative emotional response:

A: The meeting’s been cancelled . B: Yippee!

A: I’ve just dropped the box of eggs . B: Oh no!

A: I don’t think this dessert looks very fresh . B: Yuck!

Interjections ( ouch, hooray )

Word of the Day

Your browser doesn't support HTML5 audio

acting or speaking together, or at the same time

Alike and analogous (Talking about similarities, Part 1)

Learn more with +Plus

• Recent and Recommended {{#preferredDictionaries}} {{name}} {{/preferredDictionaries}}
• Definitions Clear explanations of natural written and spoken English English Learner’s Dictionary Essential British English Essential American English
• Grammar and thesaurus Usage explanations of natural written and spoken English Grammar Thesaurus
• Pronunciation British and American pronunciations with audio English Pronunciation
• English–Chinese (Simplified) Chinese (Simplified)–English
• English–Chinese (Traditional) Chinese (Traditional)–English
• English–Dutch Dutch–English
• English–French French–English
• English–German German–English
• English–Indonesian Indonesian–English
• English–Italian Italian–English
• English–Japanese Japanese–English
• English–Norwegian Norwegian–English
• English–Polish Polish–English
• English–Portuguese Portuguese–English
• English–Spanish Spanish–English
• English–Swedish Swedish–English
• Dictionary +Plus Word Lists

Add ${headword} to one of your lists below, or create a new one.

{{message}}

Something went wrong.

There was a problem sending your report.

• Join for Free

What Are Discourse Markers? Examples Explained

• by Mónica Martín Rivas @monicamartinrivas

Discover discourse markers: learn their function and use in language with clear and practical examples. Improve your communication today!

What Are Discourse Markers?

Discourse markers are words or expressions used to link sentences or paragraphs, helping to organize information and guide the reader through the text. Their function is to make the discourse more coherent and understandable, giving fluidity and structure to the content.

For example, "on the other hand", "however" or "in conclusion" are discourse markers used to indicate changes of topic, contrast or summary in a discourse. By using them effectively, texts can be made more memorable and easier for the reader to follow.

Properties and Characteristics of Discourse Markers

Discourse markers are key elements in speech writing, as they help connect ideas and give flow to the text . They are words or expressions that help organize information, show relationships between ideas and mark the structure of the discourse . They also have the ability to emphasize concepts, introduce examples or give continuity to the argument .

In speech writing, it is essential to use these markers properly in order to achieve coherence and cohesion in the message. The following are some properties and characteristics of discourse markers that will help you understand their importance in speech writing.

· Prosodic Point of View

In the study of the prosodic perspective of discourse markers, we analyze how elements of oral expression, such as stress and intonation, influence communication. Intonation is reflected in writing through the use of punctuation marks, changes in tone and the position of discourse markers. For example, rising intonation at the end of a sentence may reflect a question, while falling intonation indicates a statement.

Discourse markers may appear in different order within the utterance , which affects the interpretation and emphasis of the spoken utterance. For example, in the sentence "I was late, but not because of me", the marker "but" modifies the interpretation of the information presented.

Examples of discourse markers that reflect prosodic perspective are "furthermore," "however," "on the other hand," and "of course," which indicate changes in conversational direction and emphasize certain points. Prosody in writing and speaking is crucial for conveying the meaning and intent of speech.

· Morphological Point of View

Discourse markers are invariable and belong to different grammatical categories, such as interjections, adverbs and adverbial locutions . For example, the interjection "good" is a discourse marker that expresses approval or agreement in a discourse. As for adverbs, "too" is a discourse marker that indicates addition of information in a discourse. For their part, adverbial locutions such as "of course" function as discourse markers that reinforce the idea expressed in a discourse.

These discourse markers can affect discourse members that constitute very diverse lexical and syntagmatic categories . For example, in the sentence "well, as I told you, I would also like to add that, of course, we need to be prepared", the discourse markers "well", "also" and "of course" influence different discourse members, such as verbs, pronouns and adjectives, providing nuances of meaning and coherence in the discourse.

· Syntactic Point of View

Discourse markers are syntactic elements that are integrated into the sentence to give it a specific function within the discourse . They can appear at the beginning, in the middle or at the end of the sentence, depending on the emphasis to be given to the idea they introduce. For example, "in addition", "on the other hand" or "in summary" can appear in different positions within the sentence.

The difference between discourse markers and adverbs that function as circumstantial complements lies in the fact that the former have the function of guiding the structure of the discourse , while adverbs simply add information about the verb, adjective, or adverb they modify. Moreover, discourse markers and adverbs do not coordinate with each other and do not admit negation, since they fulfill different functions within the sentence.

· Semantic Point of View

From a semantic point of view, these markers not only indicate the relationship between the different parts of speech, but also provide nuances of attitude, modality, and discourse structure . Their referential meaning refers to their function of connecting ideas and organizing information, while their subjective meaning influences the speaker's perception and the receiver's interpretation of the message.

· Pragmatic Point of View

From a pragmatic point of view, discourse markers play a crucial role in communication, as they help to establish the relationship between the parts of speech and guide the interpretation of the message . Their correct use allows the sender to express their point of view in a clearer and more effective way, which enriches communication.

Types of Discourse Markers

Discourse markers are key elements in the construction of an effective discourse. Their proper use can make the difference between a well-structured and coherent text and a confusing and unclear one. In this sense, it is important to know the different types of discourse markers and how to use them appropriately to improve the quality of our discourse . Below, we will explain the different types of discourse markers and their importance in written and oral communication.

1. Information Structuring

These types of markers help to organize information. They are divided into:

- Commentators

Discourse markers function as commentators by introducing new comments into the discourse . Some common types include "for example", "in other words", "by the way" and "in addition". These markers introduce concrete examples to support or illustrate an idea, or rephrase previous ideas to clarify or emphasize a point . In addition, they operate as connectors between text segments to maintain discourse cohesion.

- Ordenators, Contrasters and Cause-Effect Markers

Some types of discourse markers give order to the sentence, such as "first", "then", "finally"; contrast, such as "on the other hand", "instead", or cause-effect, such as "because", "since" or "due to".

- Digressors

Digressors are connectors that are used to introduce side comments into a discourse, allowing relevant but non-essential information to be added. Examples of digressor connectors are "by the way", "having said that", "in any case" or "by the way".

For example, in a speech about the importance of caring for the environment, a digressor such as "by the way, the use of renewable energy is a viable alternative to reduce the carbon footprint" could be used.

2. Counter-Argumentative Connectors

Counter argumentative connectors are essential for linking two parts of speech so that the second acts as a suppressor or attenuator of the conclusion that could be drawn from the first . These adversative connectors introduce an idea that contradicts or limits the previous statement.

Some examples of adversative connectors are "however", "nevertheless", "in spite of" and "although". These connectors are key to identifying the beginning of an idea that presents an opposition to what has been mentioned earlier in the speech.

For example, in a speech about the impact of social networks on society, we could use an adversative connector to introduce a counterargument: "Although social networks have facilitated communication on a global level, however, they have also generated problems of addiction and social isolation".

3. Reformulators

Reformulators are a type of discourse markers whose function is to introduce a new formulation in the discourse. They are the following:

- Explanatory Reformulators

Explanatory reformulators are words or phrases that are used to clarify or expand the information presented in the previous discourse . For example, "that is" can be used to redefine an idea more clearly, "that is" to explain in other words what has already been mentioned, or "in other words" to offer a new perspective on the same information.

These types of explanatory reformulators are useful to avoid confusion and to ensure that the audience understands the message clearly and accurately .

- Rectifying Reformulators

Rectifying reformulators are discourse markers whose function is to correct or improve an incorrect formulation . In other words, they replace a first member of the discourse considered incorrect with another that improves it.

For example, by using "rather" we are indicating that we want to correct a statement or qualify an idea.

- Distancing Reformulators

Distancing reformulators are linguistic tools that allow the sender to move away from the content expressed in the discourse , either to qualify it, soften it or give it a more objective tone. Their main function is to introduce a new formulation that conditions the continuation of the discourse, while depriving the preceding discourse member of its relevance.

For example, the use of expressions such as "said differently", "better said", "in summary" or "in short" serves the function of distancing the sender from the previously expressed content, offering a new formulation in which an attempt is made to summarize or clarify the information.

- Recapitulatory Reformulators

Recapitulatory reformulators serve to present a discourse member as a conclusion or recapitulation of another member or series of members . These discourse markers help reinforce the main idea and summarize what has been said previously . For example, words such as "to summarize" or "in conclusion" indicate that a topic or argument is coming to a close.

4. Argumentative Operators

Argumentative operators are tools that condition the argumentative possibilities in a discourse . They are:

- Argumentative Reinforcement Operators

Argumentative reinforcement operators are linguistic expressions used to reinforce or emphasize an argument , giving it more force and clarity. These operators condition the argumentative possibilities of the discourse by highlighting the importance of a particular idea or argument . For example, the expression "in fact" is used to give emphasis to a statement, highlighting that what is being said is the truth or reality of the situation. Another example is the phrase "in essence," which serves to emphasize what is really intended to be conveyed, beyond what has been said superficially. Finally, the use of "in fact" is used to reinforce a statement or assertion, indicating that there is concrete evidence or data to support the argument.

- Operators of Concreteness

Operators of concreteness are linguistic tools used to give precision and specificity to a text . For example, words such as "specifically," "in particular," and "concretely" can be used to focus on a specific point within a more general discourse.

5. Conversational Markers

Conversational markers are key tools for keeping track of discourse.

- Initiating Markers

Initiating markers, such as "well, for starters," help initiate a new idea or change the topic in a conversation. For example, "Well, for starters, I want to talk about the importance of speeches in politics."

- Emphatic Markers

Emphatic markers, such as "no doubt," serve to emphasize the importance or certainty of a statement. For example: "Undoubtedly, speeches can influence public opinion".

- Approbative Markers

Approbative markers, such as "exactly", show agreement with what has just been said. For example, "Exactly, speeches are a powerful tool for persuading the audience."

- Affirmative Markers

Affirmative markers, such as "of course they do," help reinforce a previous statement. For example: "Of course, a good speech can inspire people".

- Verbative Markers

Verbative markers, such as "in fact", are used to add additional evidence or arguments to the conversation. For example, "In fact, there are studies that demonstrate the impact of speeches on society."

Understanding and correctly applying discourse markers is critical to any effective and coherent speech. But do you want to take your communication skills to the next level? Don't miss our public speaking course .

Learn more and perfect your speeches with these resources:

- Presentation Techniques: From Slide to Speech Course - Communication Techniques Course - Top public speaking courses

Recommended courses

Content Creation and Editing for Instagram Stories

A course by Mina Barrio

Discover the secrets of photography and video to find success on Instagram

• 98% ( 6.1K )

Introduction to Community Management

A course by Ana Marin

Learn to work as a community manager and build, manage and administrate a social media community

• 98% ( 4.8K )

Professional Photography for Instagram

Learn tricks to take and edit photographs for social media on your phone.

• 99% ( 11K )
• Follow Domestika

What are discourse markers and how do you use them correctly in English?

‘Discourse marker’ is a fairly formal term for a type of English vocabulary that we use all the time. These words and phrases help us to organize our ideas and clearly express our thoughts.

Discourse markers can be informal words like ‘So…’ and ‘Well…’, used mainly when speaking. They can also be more formal phrases like ‘In my opinion…’ or ‘In contrast…’, often used in essays and more formal written communication.

Here we will explain what a discourse marker is and the main types of discourse marker you are likely to need – especially if you are learning English as a second language . You may browse our pages on this topic to find lists of discourse markers for different situations, along with examples.

What are discourse markers?

Discourse markers are words and phrases which we use in written and spoken English to organize and link what we are saying. You may also hear them called ‘ transition words ‘, ‘signal words’, ‘linking words’, or even ‘fillers’.

Do discourse markers change the meaning of a sentence?

Discourse markers generally do not change the intrinsic meaning of a sentence, although sometimes they are important for clarity and logical linking of ideas.

For example, you could say:

“John has broken his leg. He attends his team’s football matches every week.”

These sentences are grammatically and factually correct, but they don’t really convey the full message. Instead, you could say:

“ Despite the fact that John has broken his leg, he still attends his team’s football matches every week.”

By adding these discourse markers, you are emphasizing John’s dedication to supporting his team despite his misfortune – without having to directly state this.

This is just one example of how discourse markers and linking phrases can add depth and clarity to your English.

Formal vs informal discourse markers in speaking vs writing

The types of discourse marker used in speaking are often quite different to those used in writing. Spoken discourse markers tend to be less formal and can include the following:

• Interjections: Oh! Wow! Ah; Ouch!
• Hesitators: Ummm; Erm; Er; Hmmm; Well
• Expletives: Damn! My God! Goodness me! Oh no! [and insert your own expletives here]
• Vocatives: Hey! Wait! Dude; Mate; Oi!

These examples of discourse markers are usually only spoken or used in informal written messages.

Of course, any linking phrases used in writing may also be spoken, but the more formal the expression, the less likely it is to be used in normal conversation.

When it comes to discourse markers in essays and formal publications, they must be carefully chosen and purposefully placed in order to guide the reader through the author’s thought process. Used well, they can also convey a more authoritative tone. However, see the warning below about overusing or misusing linking words.

Types of discourse marker

Above we gave some examples of informal discourse markers used in speech, but now let’s consider the categories of discourse marker used in writing.

Below you’ll find a short list of discourse markers for each purpose, but we have published more detailed articles on each of these topics too. If you would like to see more examples and learn how to use them properly, click through to the page indicated.

Comparing and contrasting

When you want to talk about how two things are different, or look at alternatives, you might use these words and phrases:

• On the contrary
• Nevertheless
• Even though

Read our full article about discourse markers to compare and contrast for full example sentences and a lot more useful vocabulary.

Similarities

To talk about two things that are almost the same, and highlight similarities, transition words such as these will help:

• Analagous to
• In a similar fashion
• In the same way

Often, comparison and similarity signal words come hand-in-hand when you are relating one thing to another.

These expressions are used when you need to add to what you have already said:

• As well (as)
• Additionally
• Furthermore
• What’s more
• On top of this

Read all about discourse markers for addition .

Giving examples

If you need to illustrate a point or give an example of what you mean, these phrases can help:

• For example
• For instance
• As we can see from
• As shown by
• In particular

Our article about discourse markers for giving examples has much more information.

Cause and effect / reason

When talking about one thing that caused or impacted another thing, you’ll need some of these:

• As a result of
• This has the effect of
• Consequently
• In that case

Discover more examples of discourse markers of reason and how to use them correctly. You might also want to read about affect vs effect , since these words are often confused in English.

Indicating time order

These words and phrases are helpful for giving time order or structure to a sequence of events or points:

• To begin with
• First of all
• Last of all

This is just a short list; there are more than 100 time order transition words for you to explore.

Clarification

If you need to explain something in a different way or clarify a point, you can use phrases like this:

• What I mean is
• Look at it this way
• In other words
• If I’m not mistaken
• To be clear
• This is to say that
• For the avoidance of doubt

Sometimes you need to emphasize a particular point when speaking or writing. In these cases, the following emphasis discourse markers are useful:

• As a matter of fact
• I must admit
• To be honest
• To tell you the truth
• To say the least

And finally, it can be helpful to use these words and phrases to indicate the conclusion of what we are saying:

• In conclusion
• This brings us to the conclusion that
• What this all means is
• I’ll end by saying that

For more details, you can check our full list of conclusion transition words and how to use them.

Knowing when to use or omit discourse markers

Just because all of these discourse markers exist in English, it does not mean you should use them whenever you see the opportunity.

Overuse of discourse markers can sound unnatural or overly formal, so it’s important to be able to discern when to use or omit them.

When reviewing a piece of writing, whether it’s an essay, a business case or a thought leadership piece, here are some questions you can ask yourself to check that you have used discourse markers appropriately:

• Have you used a variety of simple conjunctions (and, but, so, or…) and discourse markers?
• Have you varied the placement of these words and phrases, i.e. not always using them to begin a sentence?
• Does each paragraph serve a clear purpose?
• Have you used discourse markers to indicate the purpose without repeating them unnecessarily?
• Is there a clear link from each paragraph to the next?
• Have you used linking words and phrases to logically build up your ideas or case?

By keeping all of this in mind, you should have a good foundation for using discourse markers well.

One comment

that is useful material

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Save my name, email, and site URL in my browser for next time I post a comment.

Sign me up for the newsletter!

Discourse Markers List with Examples, Types and Uses

Discourse markers are ‘word (or phrase) tags’ that are used at the beginning of sentences (few times they can even be used in the middle) to help them seem clearer and more understandable. They help join ideas and act as a connection or link between sentences and paragraphs.

Discourse markers are effective in both speeches as well as in writing. They help give a conversation or a piece of text a proper shape and assist in making it appear more fluent and logical. They allow the writer or speaker to use them as ‘fillers’.

Since they are ‘word tags’ and act only as a link between two thoughts, they are separated from the main sentence by a comma. Therefore, a sentence still retains its structure even if the discourse marker is removed from the beginning of the sentence.

Different discourse markers are used in different situations. It is very important to use them in the right situation or they will seem completely out of place. Given below are a few common discourse markers used in everyday English. It is important to learn their correct usage.

Table of Contents

Discourse Markers List

1. Addition

• Furthermore
• In addition
• Additionally
• Not only… but also

2. Contrast

• On the other hand
• Nevertheless
• Even though
• Nonetheless

3. Cause and Effect

• Consequently
• As a result

4. Emphasis

• Particularly
• Undoubtedly
• Importantly

5. Clarification

• In other words
• To put it another way

6. Illustration

• For example
• For instance
• Specifically

7. Sequence

• Subsequently
• Simultaneously

8. Conclusion

• In conclusion
• To conclude
• In my opinion
• It seems to me
• From my perspective
• As far as I’m concerned
• By the time

11. Condition

• Provided that
• On condition that

12. Agreement

• By all means

13. Disagreement

• On the contrary
• In contrast

14. Limitation

15. Possibility

• Potentially

Discourse Markers Used in Sentences

• And : I love playing the guitar, and I also enjoy the piano.
• Also : She’s a talented artist; also , she writes beautifully.
• Plus : I’ve finished my homework; plus , I’ve even done some extra reading.
• Furthermore : He’s a great team player; furthermore , he always takes the initiative.
• Moreover : She’s an excellent leader; moreover , she has a vision for the future.
• In addition : I like reading fiction; in addition , I have a keen interest in poetry.
• Additionally : I enjoy hiking; additionally , I’ve started bird watching recently.
• Not only… but also : He’s not only a skilled painter but also a sculptor.
• Besides : I can play the guitar; besides , I’m learning the violin.
• But : I wanted to go for a walk, but it started raining.
• Yet : I was tired, yet I finished the assignment.
• However : I love chocolate; however , I must limit my intake for health reasons.
• On the other hand : I enjoy reading fiction; on the other hand , my brother loves non-fiction.
• Conversely : Some people love the city life; conversely , others prefer the countryside.
• Nevertheless : It was raining; nevertheless , we went for a hike.
• Although : Although I’m scared of heights, I tried zip-lining.
• Even though : Even though he’s younger, he’s quite mature for his age.
• While : While I love coffee , my sister prefers tea.
• Whereas : I enjoy action movies, whereas my friend loves romantic films.
• So : I was hungry, so I made a sandwich.
• Because : I went to bed early because I was tired.
• Since : Since it’s raining, we should stay indoors.
• Therefore : He didn’t study; therefore , he failed the exam.
• Consequently : She forgot her umbrella; consequently , she got wet in the rain .
• As a result : The road was slippery; as a result , driving was dangerous.
• Hence : He was late, hence he missed the beginning of the movie.
• Thus : She practiced daily; thus , she became a great pianist.
• Due to : The match was canceled due to heavy rain.
• Owing to : The flight was delayed owing to technical issues.
• Indeed : The artwork is beautiful; indeed , it’s one of the best I’ve seen.
• In fact : He’s not just a good player; in fact , he’s the team captain.
• Especially : I love all fruits, especially mangoes.
• Particularly : The concert was good, particularly the last performance.
• Clearly : Clearly , there’s been a misunderstanding.
• Obviously : Obviously , we need to reconsider our strategy.
• Undoubtedly : Undoubtedly , this is a significant achievement.
• Importantly : Importantly , always remember to be kind.
• I mean : I love aquatic animals, I mean , creatures that live in water.
• In other words : The weather is inclement; in other words , it’s unpredictable and severe.
• That is : I prefer a sedentary lifestyle, that is , I like staying in one place.
• Namely : I have two favorite fruits, namely , apples and bananas.
• To put it another way : He’s frugal; to put it another way , he’s careful with money.
• To clarify : To clarify , I was referring to the meeting next week, not tomorrow.
• For example : There are many yoga poses; for example , the downward dog and the cobra.
• For instance : I love reading classics; for instance , works by Dickens and Austen.
• Such as : I enjoy various sports, such as basketball, tennis, and swimming.
• Like : I have hobbies like painting and gardening.
• Namely : I have three pets, namely , a cat, a dog, and a parrot.
• Specifically : I’m interested in ancient civilizations, specifically the Egyptians and the Mayans.
• First : First , preheat the oven to 350°F.
• Next : Next , mix the dry ingredients in a bowl.
• Then : Then , add the wet ingredients and stir.
• Lastly : Lastly , pour the batter into a pan and bake.
• Finally : Finally , let the cake cool before serving.
• After that : I’ll go to the gym; after that , I’ll pick up some groceries.
• Subsequently : He completed his degree; subsequently , he secured a job in a reputed company.
• Meanwhile : I’ll prepare the sauce; meanwhile , can you boil the pasta?
• Simultaneously : The orchestra played simultaneously with the choir.
• Thereafter : He graduated and thereafter pursued higher studies.
• In conclusion : In conclusion , a balanced diet and regular exercise are key to good health.
• To sum up : To sum up , the event was a huge success.
• In summary : In summary , the research supports the hypothesis.
• All in all : All in all , it was a memorable trip.
• Overall : Overall , the feedback was positive.
• To conclude : To conclude , further studies are needed to confirm these findings.
• In short : In short , practice makes perfect.
• In my opinion : In my opinion , the book is worth reading.
• I believe : I believe everyone should have access to education.
• I think : I think it’s going to rain today.
• Personally : Personally , I prefer tea over coffee.
• It seems to me : It seems to me that he’s not interested in
• Now : Now , let’s move on to the next topic.
• Today : Today , we’ll be discussing environmental issues.
• Recently : Recently , there has been a surge in online shopping.
• Previously : Previously , this building was a school.
• Soon : The movie will start soon .
• Later : We can discuss this later .
• Until : Wait here until I return.
• Whenever : Call me whenever you need help.
• As long as : You can stay as long as you like.
• By the time : By the time we arrived, the show had already started.
• If : If it rains, the match will be canceled.
• Unless : I won’t go unless you come with me.
• Provided that : You can join the club provided that you pay the membership fee.
• In case : Take an umbrella in case it rains.
• As long as : As long as you’re happy, that’s what matters.
• Supposing : Supposing we miss the train, what’s our backup plan?
• On condition that : You can borrow my book on condition that you return it next week.
• Of course : Would you like some tea? – Of course !
• Definitely : Are you coming to the party? – Definitely !
• Surely : This is, surely , the best cake I’ve ever tasted.
• Indeed : The concert was fantastic, indeed .
• Naturally : Naturally , parents are concerned about their children’s safety.
• By all means : Can I use your phone? – By all means .
• However : I like the design; however , I’m not sure about the color.
• On the contrary : It’s not expensive; on the contrary , it’s quite affordable.
• In contrast : Summer is hot and sunny; in contrast , winter is cold and snowy.
• Nonetheless : It was raining; nonetheless , we decided to go out.
• But : I wanted to buy the dress, but it was too expensive.
• Yet : I tried my best, yet I couldn’t complete the task.
• Although : Although she was tired, she continued working.
• Even though : Even though he had little experience, he managed to do a great job.
• However : I love the concept; however , the execution could be better.
• Maybe : Maybe I’ll join you later.
• Perhaps : Perhaps we can find a solution together.
• Possibly : It’s possibly the best option we have.
• Potentially : This is a potentially groundbreaking discovery.
• Could : We could go to the movies tonight.
• Might : She might come to the party if she’s free.

Here is an example of how a few discourse markers can be used in writing:

Once upon a time, there was a boy called Tom. He lived on a hill and picked berries for a living. He would save some berries for himself and his family and sell the rest to a fruit seller in a nearby town. The fruit seller was very happy with Tom because he would bring him a wide variety of berries. For instance, he would bring him strawberries, cherries, raspberries, blackberries, blueberries and mulberries. Furthermore, he would throw away the rotten ones and wash and clean the good ones before giving them to the fruit seller. Therefore, as a reward the fruit seller would give Tom one dozen bananas and mangoes each month for free.

As Tom grew older he grew tired of picking berries. He wanted to become a woodcutter like his father. However, his father insisted that Tom continued to do his old work for some time. Tom had two elder brothers, Jack and Mark. Jack was a cobbler in the town, whereas Mark made bread in a bakery there. Similarly, Tom’s mother also worked as a seamstress at the tailoring shop in the town.

Days went by as Tom’s urge to pick up an axe grew stronger. One morning Tom hid behind his house and waited for his father, mother and brothers to leave for work. When everyone left, he went inside, got his father’s spare axe and walked into the woods. He came upon a small tree which he thought would be easy to cut, but just as he swung the axe, it flew from his hands and hit a bird that was perched on the lower branch of the tree. The bird was badly injured and started to bleed. Tom froze in shock at what had happened. Meanwhile, his father, who had forgotten his lunch at home and was walking back to get it, saw him standing like a statue in the nearby woods. He approached the scene and quickly assessed what had happened. He took out his handkerchief, wrapped the bird in it and rushed it to the stable, where the town veterinarian worked.

When Tom’s father returned home that evening, he was quiet but upset. He couldn’t believe his son had disobeyed him like this. Nonetheless , he called Tom before going to bed and explained why he had not allowed him to cut wood. He told him that he was neither strong enough nor ready for such a responsibility at that age. He promised that if Tom was patient for just a few more years, he would teach him to cut wood himself.

In the end, Tom reflected on his actions as he went to sleep that night and decided that he would rather wait for a while and be his father’s woodcutting partner than be hasty and hurt his father, as well as the animals in the woods.

Get PDF Here

You May Also Like

• Dry Fruits Names
• Fruits Names in English
• 500 Adverbs List A To Z
• Linking Words
• Gerunds in English

discourse markers discourse markers activity Discourse Markers Examples discourse markers exercises Discourse Markers in English discourse markers lesson plan list of discourse markers text with discourse markers

Determiners in English with Types & Examples

Pronouns Words, Definition Kinds with Examples

Copyright © 2024 by englishan

Username or Email Address

Remember Me

Forgot password?

Enter your account data and we will send you a link to reset your password.

Your password reset link appears to be invalid or expired.

Privacy policy.

To use social login you have to agree with the storage and handling of your data by this website. %privacy_policy%

Add to Collection

Public collection title

Private collection title

No Collections

Here you'll find all collections you've created before.

The English Bureau

English for IELTS, Business & Advanced

DISCOURSE MARKERS – LINKING WORDS

30th October 2019 By Alex Markham 11 Comments

Discourse markers are words or expressions that link, manage and help to organise sentences. They connect what is written or said with something else. They make no change to the meaning.

They are also often called linking words and, sometimes, fillers . They are important to make your speech or text flow and to avoid a series of short unconnected statements.

USING DISCOURSE MARKERS

This post contains affiliate links. If you buy through these links, the price you pay will be the same as if you go direct.

Discourse markers can be placed in any part of a sentence, including the beginning. We use different types of discourse markers for different types of links; informal markers for speech and formal type markers for formal writing, such as essays and reports.

Discourse markers do not always have meanings that you will find in your dictionary. Instead, they provide certain functions such as delaying, filling or hinting at emotions.

Discourse markers are important for fluency in English at an advanced level. They are important elements of speech and writing for living and working in English-speaking countries.

If you need to take IELTS exams, you’ll need to have a clear understanding of discourse markers for the Writing and Speaking tests.

INFORMAL AND SPOKEN DISCOURSE MARKERS

When speaking or informal writing, such as emails or texts to friends, the following discourse markers are often used. There are many of course, but here are some of the most common.

• Anyway – marks a shift away from a topic. It’s sold out, anyway, I didn’t want to go .
• Actually – indicates that what you are saying is a surprise to you or is the opposite of what might be imagined or introduces the opposite response to what someone wants or expects. I thought I didn’t like dance shows but actually , I quite enjoyed it. Actually, I’d prefer if if you didn’t smoke in here.
• As I was saying / As I said – continuing a theme after a change of subject. As I was saying before I was interrupted, I’d like to go to the cinema . As I said last week, I don’t enjoy opera.
• Absolutely – used as a stronger way of saying yes. A. Would you like a holiday in Thailand? B. Absolutely , when do we go?
• Exactly – used to emphasise what you mean. What exactly are you talking about? Exactly , that’s what I meant.
• Okay or OK – understood or agreed. Okay , I get it.
• Well – used to indicate an emotion including annoyance, anger, surprise, disappointment and relief. Well , that’s not what I wanted to hear. Well , I didn’t expect that!
• Fine – used informally to indicate that something is not good. That’s another fine mess you’ve got me into. A. How are you since your wife left you? B, Fine , what do you think
• Mind (you) – used to highlight or stress something you said to avoid misunderstanding. He couldn’t help me, mind you , I know he’s busy so it’s not his fault. Or to introduce something else that needs to be considered. My kids never listen to me. Mind you, I didn’t either at their age so I guess it’s normal.
• You know – indicates that what you’re saying is known by the listener. or to check they understand with a question mark. Also used as a filler while you’re thinking of what to say next. Studying economics is hard, you know.
• I mean – introduces a statement to justify what you’ve said. It was a tough driving test, I mean I took my test in the rush hour in the rain.
• For a start – introduces the first item of a logical sequence or the most important item of other possibilities. The driving test in Spain is difficult, for a start it’s all in Spanish.
• What’s more – introduces a new fact or argument and implies that this new fact or argument is more important than the one just given. Used without contraction for formal speech and writing. The stock market fell in 2008 and what’s more , our pension fund was tied up in stocks .
• On top of that – used to indicate something unpleasant in addition to what you’ve just said. I lost my job and on top of that, I was fined for speeding.
• To tell you the truth/truth be told – indicates that what you’re saying is honest and open. To tell you the truth , I was more than a little worried. Truth be told , I’m concerned about the project
• To be honest – explains that you are being truthful about what you’re saying, usually with an understanding that the person you’re talking to will be disappointed. To be honest , I don’t think your offer is sufficient.
• Sort of / kind of – used to describe that something is partly true but it’s not the entire explanation. Or that something is difficult to explain entirely. Your answer is sort of correct but there’s a bit more to it than that. My new car is a kind of blue-green colour . Not sure what colour you’d call it.

FORMAL AND WRITTEN DISCOURSE MARKERS

Discourse markers for more formal writing situations organise your text logically. They are typical when writing formal letters or when writing essays or reports, especially at work.

They will also be used in formal speaking situations, such as presentations or speeches. For the IELTS writing exam, they are necessary to gain maximum marks under coherence and cohesion which is 25% of the mark.

• Firstly, secondly and thirdly – organise your points logically. Firstly we need to reduce the prices of the products. Secondly , we need to upgrade the channel to market etc. etc.
• Finally – marks the final point of a list. Finally, I want to talk about the environment.
• To begin with –
• In addition – provides extra information. In addition to my earlier point, we also need to think about delivery.
• In conclusion – marks the summary and round-up of your essay or speech. In conclusion , I have demonstrated that savings will be 10%.
• In summary – another way to mark the conclusion. In summary , the average costs will reduce.

• Moreover – as a further matter, there is even more to tell you. It’s a difficult decision, moreover , they would have to move home.
• On (the) one hand, on the other hand – introduces two opposing ideas. On the one hand , it’s cheaper, on the other hand, it means a lot of work to make it worth the trouble.
• What is more – introduces a new fact or argument and implies that this new fact or argument is more important than the one just given. Used with contraction for informal. The stock market fell in 2008, what is more, is that the company had all its assets in stocks.
• Furthermore – introduces an additional fact or argument. Unemployment rates fell in 2018, furthermore , average salaries rose.
• Not only….but also – used to give emphasis to a second point. This phrase uses inversions. I not only worked in London but also in Paris.
• For example – used to give an example to show that something is true. This will never work, for example the spring is broken.
• Such as – introduces an example or examples of what you’re talking about. There are lots of types of car on the road today, such as saloons, hatchbacks and four-by-fours.
• As a result – introduces what happened because of something. He failed his exams and, as a result , couldn’t attend university.
• Consequently – because of this or as a result. The road is closed, consequently we’ll need to wait until it’s open again.
• Nevertheless – despite what has just been said or written. There are insufficient funds for the development, nevertheless, we’ll continue the work and hope something turns up.
• By/in contrast – a comparison which introduces something opposite to the first item. The economy in the EU shrunk by 2%, by contrast , the USA saw growth of nearly 5%.
• In comparison – used to discuss the ways in which two things are different. Costs have risen in comparison to last year.
• In my opinion – Used to give your opinion or point of view when you are confident of the facts. In my opinion , it’s better to invest in language training as early as possible in our schools .
• Of course – stating something that is known or obvious but it’s important to mention anyway. Of course, children benefit greatly from playing sports.

ORGANISING AND LINKING SENTENCES

A discourse marker is a word or an expression that organises our sentences. Also known as linking words, they can be thought of as the glue that connects sentences together.

Formal discourse makers are vital in academic and business texts as they organise the sequencing and connections.

Of course, there are many more discourse markers in use than those I’ve listed in this article. These ones should serve to cover most situations and you will pick up even more as you use English regularly at an advanced level

Click on the cover below to find out more information on how to pass the IELTS Writing Task

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

2nd October 2020 at 20:26

Hi, Your web content is absolutely helpful for any one learning English language and preparing for IELTS. I am truly amazed and very to luckily come over your document. Thanks a lot.

Best Regards Pradeep

3rd October 2020 at 20:12

I’m glad it helps Pradeep

23rd October 2020 at 00:08

Great text! Very useful for my personal use and teaching. Thank you, Alex.

23rd October 2020 at 00:24

Hello Elder, I’m pleased it helps. Any questions, let me know

4th November 2021 at 23:20

Hi Alex. After a year. I find myself planning a lesson on discourse markers again and I wonder if you could help me with some common examples of discourse markers used for deduction.

28th April 2021 at 16:07

Hi, I just wanna say this website was so useful for me for doing my assignment ! Thank youuuuuuuuuuu !!!

2nd May 2021 at 19:15

Excellent, I’m pleased it helped

29th May 2021 at 15:33

Great content. I am studying IELTS and these words are so important in both writing and speaking. I have learnt a lot from this page. Thank you for helping me and people like me. God bless you.

3rd June 2021 at 19:21

I’m pleased my website has helped you Abhishek

6th February 2023 at 08:05

Really I love this lesson.

6th February 2023 at 14:54

Leave a comment Cancel reply

This site uses Akismet to reduce spam. Learn how your comment data is processed .

All content on the site, including personal photographs, is © Alexander Markham 2019

Privacy Policy

• April 14, 2022

What are speech marks? Examples and how to use them

From novels to news reports, speech marks are used in all sorts of places! In this guide, we take a look at what they are, where they’re used and some handy rules to keep in mind when using them.

In this blog, jump to:

What are speech marks?

Where are speech marks used, when do children learn about speech marks, how to use speech marks, speech mark examples, how to use speech marks when writing, general rules for using speech marks.

Speech marks are special punctuation marks that show the exact words someone has spoken.

They’re also known as quotation marks and inverted commas, and they always appear in pairs at the beginning and end of what was said.

This is what they look like: “ ”

And here’s an example of some speech marks in action:

“The words that appear between speech marks are the exact words that someone has said,” said the teacher.

You can find speech marks everywhere. They’re common in stories, news articles, poetry and essays. By showing when someone is speaking, speech marks can help make a text more lively and easier to read.

In a story, speech marks may show which of the characters is speaking. For example, in a children’s book you may read:

“We always eat in the garden,” said Markus.

In a news article, speech marks can be used to show a statement made by an expert. For example:

Economists say that “the increase in production will help the job market to grow”.

Children usually start learning about speech marks around the age of 7.

Before they learn about speech marks, they’ll first be taught about other forms of punctuation, including full stops, commas, exclamation marks and question marks.

Being able to tell the difference between direct and indirect speech will help children to effectively use speech marks.

• Direct speech: shows the exact words that were spoken by a person or character. Speech marks are used at the beginning and end of direct speech to represent what someone has said.
• Indirect speech: shows what someone said without using their exact words. Speech marks aren’t used in these cases.
• A reporting clause will let a reader know who said the direct speech and how they said it.

You should always separate direct speech from the reporting clause by a comma.

This may sound a little complicated, but if we look at an example it will all be clearer.

Anne said, “It’s snowing!”

In this example, ‘Anne said’ is the reporting clause, as it lets us know that it was Anne who spoke. This clause is separated by a comma from Anne’s direct speech, which is represented by the text between the speech marks (“It’s snowing!”).

If you’re writing a conversation between multiple people, reporting clauses can also help to clarify who’s speaking, and when.

Let’s take a look at some examples of speech marks in actions. These will help us better understand the different uses of speech marks.

The teacher said that Anthony’s story was “exceptionally written” and that’s why she read it to the whole class.

In this example, a fragmented part of the teacher’s statement is quoted. This is why the statement isn’t capitalised and has no punctuation before or within the quotation marks.

Markus laughed, “Did you see how high I jumped?”

“I know! It was amazing!” Adriana agreed.

In this example, two different people are speaking. The reporting clauses (‘Marcus laughed’ and ‘Adriana agreed’) shows us who is speaking. Each speaker is also placed on a new line to help the reader see that a new person has started speaking.

“It’s amazing,” William exclaimed, “there are bubbles everywhere!”

In this example, the statement made by William is separated by the reporting clause. This is a fun way of adding some character to your writing when using speech marks.

“Polar bears are my favourite real animals,” Mario said. He then told us about the mythical animals he likes more than polar bears.

This is an interesting example because only part of what Mario said is directly written down. The rest is summarised in indirect speech, and this is why there are no speech marks in the second sentence.

If you’d like to quote a phrase or statement, using speech marks will show the reader that these aren’t your original words.

Speech marks are also a great way to give emphasis to certain phrases.

If you’re writing stories, they’ll also help to separate when the different characters are talking. Keep in mind that speech is a faster way of moving a story forward, so be careful about how much you use them!

There are a few helpful rules to keep in mind when using speech marks:

• Sentences made by different characters or people will need to be in separate lines
• If you’re using a person’s full statement, the punctuation of the sentence will need to be within the quotation marks. This includes full stops, exclamation marks, and question marks.
• If a reporting clause is before the speech marks, you’ll need a comma to separate the two. This comma will be before the speech marks.
• If a reporting clause comes in the middle of two quoted sentence parts, you’ll need a comma at the end of the speech marks in the first part and at the beginning of the speech marks in the second part of the sentence

Final thoughts

Using speech marks is relatively easy once you realise that they’re only used to express what someone said. When writing an essay or story, speech marks can also help to make your writing more interesting and even livelier!

To learn even more about speech marks, why not download the DoodleEnglish app ?

It’s filled with interactive exercises exploring the topic, making it the perfect way to put your learning into practice.

Best of all, it’s designed to be used for just 10 minutes a day — and you can try it for absolutely free!

Related posts

What we offer

Quick links

All rights reserved. 

Are you a parent, teacher or student?

Get started for free!

Maths information pack

We ask for your contact info so we can send our info pack directly to your inbox for your convenience, exam prep information pack, case studies information pack.

Book a chat with our team

I’m new to Doodle

My school is already using Doodle

Information pack

We ask for your contact info so that our education consultants can get in touch with you and let you know a bit more about doodle., student login, which programme would you like to use.

DoodleMaths

DoodleTables

DoodleEnglish

DoodleSpell

If you’d like to use Doodle’s browser version, please visit this page on a desktop.

To log in to Doodle on this device, you can do so through our apps. You can find out how to download them here:

Linking Your Ideas in English With Discourse Markers

Hero Images/ Getty Images

• Pronunciation & Conversation
• Writing Skills
• Reading Comprehension
• Business English
• Resources for Teachers
• TESOL Diploma, Trinity College London
• M.A., Music Performance, Cologne University of Music
• B.A., Vocal Performance, Eastman School of Music

Some words and phrases help to develop ideas and relate them to one another. These kinds of words and phrases are often called discourse markers . Note that most of these discourse markers are formal and used when speaking in a formal context or when presenting complicated information in writing.

with regard to / regarding / as regards / as far as ……… is concerned / as for

These expressions focus attention on what follows in the sentence. This is done by announcing the subject in advance. These expressions are often used to indicate a change of subject during conversations. 

His grades in science subjects are excellent. As regards humanities … With regard to the latest market figures we can see that ... Regarding our efforts to improve the local economy, we have made ... As far as I am concerned, we should continue to develop our resources. As for John's thoughts, let's take a look at this report he sent me.

on the other hand / while / whereas

These expressions give expression to two ideas which contrast but do not contradict each other. 'While' and 'whereas' can be used as subordinating conjunctions to introduce contrasting information. 'On the other hand' should be used as an introductory phrase of a new sentence connecting information.

Football is popular in England, while in Australia they prefer cricket. We've been steadily improving our customer service center. On the other hand, our shipping department needs to be redesigned. Jack thinks we're ready to begin whereas Tom things we still need to wait.

however / nonetheless / nevertheless

All these words are used to begin a new sentence which  contrasts two ideas . These words are often used to show the something is true despite not being a good idea. 

Smoking is proved to be dangerous to the health. Nonetheless, 40% of the population smokes. Our teacher promised to take us on a field trip . However, he changed his mind last week. Peter was warned not to invest all of his savings in the stock market. Nevertheless, he invested and lost everything.

moreover / furthermore / in addition

We use these expressions to add information to what has been said. The usage of these words is much more elegant than just making a list or using the conjunction 'and'.

His problems with his parents are extremely frustrating. Moreover, there seems to be no easy solution to them. I assured him that I would come to his presentation. Furthermore, I also invited a number of important representatives from the local chamber of commerce. Our energy bills have been increasing steadily. In addition to these costs, our telephone costs have doubled over the past six months.

therefore / as a result / consequently

These expressions show that the second statement follows logically from the first statement.

He reduced the amount of time studying for his final exams . As a result, his marks were rather low. We've lost over 3,000 customers over the past six months. Consequently, we have been forced to cut back our advertising budget. The government has drastically reduced its spending. Therefore, a number of programs have been canceled.

Check our understanding of these discourse markers with this short quiz. Provide an appropriate discourse marker in the gap. 

• We've done a great job on the grammar. ______________ listening, I'm afraid we still have some work to do.
• __________ Americans tend to eat quickly and leave the table, Italians prefer to linger over their food. 
• The company will introduce three new models next spring. __________, they expect profit to rise considerably. 
• He was excited to go to the movies. ____________, he knew that he needed to finish studying for an important exam.
• She warned him repeatedly not to believe everything he said. __________, he continued believing him until he found out that he was a compulsive liar. 
• We need to consider every angle before we begin. _________, we should speak with a number of consultants on the matter. 
• With regard to / Regarding / As regards / As for
• while / whereas
• Therefore / As a result / Consequently
• However / Nonetheless / Nevertheless
• On the other hand
• In addition / Moreover / Furthermore
• Comparing and Contrasting in English
• Text Organization
• How to Use Sentence Connectors to Express Complex Ideas
• Showing Addition in Written English
• Showing Cause / Effect in Written English
• How to Use the Preposition 'To'
• Sentence Connectors: Showing Opposition in Written English
• Uses for the Preposition "At"
• How to Use Just in English
• The Right Way to Use 'Already' and 'Yet' in English
• Using Adverb Clauses with Time Expressions
• Telling Stories: Sequencing for ESL Students
• How to Discuss Charts and Graphs in English
• Adverb Placement in English
• A Short Guide to Punctuation
• Proper Usage of "That" in English

Discourse Markers

Discourse markers (words like 'however', 'although' and 'Nevertheless') are referred to more commonly as 'linking words' and 'linking phrases', or 'sentence connectors'. They may be described as the 'glue' that binds together a piece of writing, making the different parts of the text 'stick together'. They are used less frequently in speech, unless the speech is very formal.

Without sufficient discourse markers in a piece of writing, a text would not seem logically constructed and the connections between the different sentences and paragraphs would not be obvious.

Care must also be taken, however, to avoid over-use of discourse markers. Using too many of them, or using them unnecessarily, can make a piece of writing sound too heavy and 'artificial'. They are important, but must only be used when necessary.

  Key advice

1. What are the different discourse markers that can be used?

2. How can sentence connectors be replaced in order to increase variety in writing?

3. How are paragraphs linked together?

4. Discourse markers in a sample passage of academic English

  What are the different discourse markers that can be used?

There are many discourse markers that express different relationships between ideas. The most common types of relationship between ideas, and the sentence connectors that are most often used to express these relationships, are given in the table below. The discourse markers in the table are generally used at the start of a phrase or clause . (a clause is a minimal grammatical structure that has meaning in its own right, and consists of a subject and verb, and often an object too). Sentence connectors do not always begin a completely new sentence; they may be separated from the previous idea with a semi-colon.

Note that there are two particular features of the sentence connectors indicated below:

• Sentence connectors can be used to begin a new sentence or a new clause that follows a semi-colon;
• Some sentence connectors can be placed in different positions within the sentence: initial position (e.g. Because he is ill, he needs to rest.) and 'mid-way position' at the start of another clause (e.g. He must rest, because he is ill).

How can sentence connectors be replaced in order to increase variety in writing?

In your writing, you will want to spend some time ensuring that your work has a sense of variety. In order to do this, you might think of the following:

• Use conjunctions as well as/instead of sentence connectors . A conjunction is a word like 'and', 'but', etc, which is used to join two ideas together into a complex sentence. Unlike sentence connectors such as 'However', etc, a conjunction cannot be used at the beginning of a sentence and must come at a mid-point, at the end of one clause and the beginning of another. It is usually possible to rephrase a pair of sentences that use a sentence connector by using a conjunction instead. For example, instead of saying 'He studied French; however, his wife studied Physics', it might actually be more natural to say 'He studied English but his wife studied Physics'. Similarly, instead of saying 'English is hard; therefore, one must spend a lot of time practising it', we can say: 'English is hard so one must spend a lot of time practising it.' These are simple examples, but the principle of paraphrase can be extended to other, more complex sentences.
• Use conjunctions at least some of the time . Words like and and but may seem boring, but they help to lighten the style of your writing. This in turn helps the writing to sound less pompous and formal. And in any case, in writing, it is often helpful to use a variety of structures rather than just saying things in one way.
• It can also be helpful to omit discourse markers if they do not serve any useful purpose. Knowing when to omit the discourse marker is a subtle aspect of language use and comes with more practice and wider reading.
• Try joining two clauses together by making one subordinat e to the other. If we go back to the sentence 'He studied English but his wife studied Physics', we can rephrase this as follows: 'He studied English whereas his wife studied Physics', or 'He studied English while his wife studied Physics.' The clause beginning with while/whereas is subordinate. this means that it is used to qualify/add extra information to the sentence, but cannot stand on its own.
• Remember, it can be tedious to read a piece of writing which has too many discourse markers. The writing can seem pedantic, heavy and over-pompous. You are ideally seeking a light, flowing style, not a heavy or forced one.

  back ^

How are paragraphs linked together?

In much the same way that ideas within a paragraph are linked, a new paragraph must be linked in some way with the previous one. This, too, necessitates the use of discourse markers.

In the table below are some different ways in which the opening of a paragraph can link back to what has happened before. The three basic types of paragraph-paragraph relationship are : reinforcement of idea; contrast of ideas; and concession. Indicating these relationships builds a 'bridge' between paragraphs and makes reading the text easier.

In your own writing, it is useful to consider the following points:

• Do my paragraphs serve a definite purpose?
• What is the exact link between the paragraph I have written and the previous paragraph? Is that link clear?
• Are my paragraphs laid out strategically , in order to help me to organise my materials to best effect?
• Do my paragraphs help me to build up my ideas in a logical and gradual fashion?

China National Conference on Chinese Computational Linguistics

CCL 2022: Chinese Computational Linguistics pp 3–16 Cite as

Discourse Markers as the Classificatory Factors of Speech Acts

• Da Qi   ORCID: orcid.org/0000-0001-5690-9799 14 ,
• Chenliang Zhou   ORCID: orcid.org/0000-0002-6574-1454 14 &
• Haitao Liu   ORCID: orcid.org/0000-0003-1724-4418 14 , 15  
• Conference paper
• First Online: 06 October 2022

495 Accesses

Part of the book series: Lecture Notes in Computer Science ((LNAI,volume 13603))

Since the debut of the speech act theory, the classification standards of speech acts have been in dispute. Traditional abstract taxonomies seem insufficient to meet the needs of artificial intelligence for identifying and even understanding speech acts. To facilitate the automatic identification of the communicative intentions in human dialogs, scholars have tried some data-driven methods based on speech-act annotated corpora. However, few studies have objectively evaluated those classification schemes. In this regard, the current study applied the frequencies of the eleven discourse markers ( oh, well, and, but, or, so, because, now, then, I mean , and you know ) proposed by Schiffrin [ 24 ] to investigate whether they can be effective indicators of speech act variations. The results showed that the five speech acts of Agreement can be well classified in terms of their functions by the frequencies of discourse markers. Moreover, it was found that the discourse markers well and oh are rather efficacious in differentiating distinct speech acts. This paper indicates that quantitative indexes can reflect the characteristics of human speech acts, and more objective and data-based classification schemes might be achieved based on these metrics.

This is a preview of subscription content, log in via an institution .

Buying options

• Available as PDF
• Read on any device
• Instant download
• Own it forever
• Available as EPUB and PDF
• Compact, lightweight edition
• Dispatched in 3 to 5 business days
• Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

For all the clustering results using different methods and distance metrics, see Appendix A .

Allen, J., Core, M.: Draft of DAMSL: dialog act markup in several layers (1997). http://www.fb10.uni-bremen.de/anglistik/ling/ss07/discourse-materials/DAMSL97.pdf

Anthony, L.: Antconc 4.0.5. Waseda University (2021)

Google Scholar  

Ariel, M.: Discourse markers and form-function correlations. In: Jucker, A.H., Ziv, Y. (eds.) Discourse Markers: Descriptions and Theory, pp. 223–260. John Benjamins, Philadelphia (1998)

Chapter   Google Scholar  

Carletta, J.: Assessing agreement on classification tasks: the kappa statistic. Comput. Linguist. 22 (2), 249–254 (1996)

Carletta, J., Isard, A., Isard, S., Kowtko, J.C., Doherty-Sneddon, G., Anderson, A.H.: The reliability of a dialogue structure coding scheme. Comput. Linguist. 23 (1), 13–31 (1997)

D’Andrade, R.G., Wish, M.: Speech act theory in quantitative research on interpersonal behavior. Discourse Process. 8 (2), 229–259 (1985). https://doi.org/10.1080/01638538509544615

Article   Google Scholar  

Fischer, K.: Frames, constructions, and invariant meanings: the functional polysemy of discourse particles. In: Fischer, K. (ed.) Approaches to Discourse Particles, pp. 427–447. Elsevier, Oxford (2006)

Fraser, B.: Pragmatic markers. Pragmatics 6 (2), 167–190 (1996). https://doi.org/10.1075/prag.6.2.03fra

Furkó, P.B.: Discourse Markers and Beyond: Descriptive and Critical Perspectives on Discourse-Pragmatic Devices Across Genres and Languages. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-37763-2

Book   Google Scholar  

Hewitt, L.E., Duchan, J.F., Segal, E.M.: Structure and function of verbal conflicts among adults with mental retardation. Discourse Process. 16 (4), 525–543 (1993). https://doi.org/10.1080/01638539309544852

Holtgraves, T.: The production and perception of implicit performatives. J. Pragmat. 37 (12), 2024–2043 (2005). https://doi.org/10.1016/j.pragma.2005.03.005

Jacobs, S., Jackson, S.: Argument as a natural category: the routine grounds for arguing in conversation. W. Jo. Speech Commun. 45 (2), 118–132 (1981). https://doi.org/10.1080/10570318109374035

Jucker, A.H.: The discourse marker well: a relevance-theoretical account. J. Pragmat. 19 (5), 435–452 (1993). https://doi.org/10.1016/0378-2166(93)90004-9

Jurafsky, D., Shriberg, L., Biasca, D.: Switchboard SWBD-DAMSL shallow-discourse-function annotation coders manual, draft 13 (1997). https://web.stanford.edu/~jurafsky/ws97/manual.august1.html

Labov, W., Fanshel, D.: Therapeutic Discourse: Psychotherapy as Conversation. Academic Press, New York (1977)

Liu, S.: An experimental study of the classification and recognition of Chinese speech acts. J. Pragmat. 43 (6), 1801–1817 (2011). https://doi.org/10.1016/j.pragma.2010.10.031

Matei, M.: Discourse markers as functional elements. Bull. Transilvania Univ. Braşov 3 (52), 119–126 (2010)

McHugh, M.L.: Interrater reliability: the kappa statistic. Biochemia Med. 22 (3), 276–282 (2012)

Article   MathSciNet   Google Scholar  

Mulkay, M.: Agreement and disagreement in conversations and letters. Text - Interdisc. J. Study Discourse 5 (3), 201–228 (1985). https://doi.org/10.1515/text.1.1985.5.3.201

Potts, C.: Switchboard Dialog Act Corpus with Penn Treebank links (2022). https://github.com/cgpotts/swda

Redeker, G.: Discourse markers as attentional cues at discourse transitions running head: discourse transitions. In: Fischer, K. (ed.) Approaches to Discourse Particles, pp. 339–358. Brill, Leiden, The Netherlands (2006). https://doi.org/10.1163/9780080461588_019

Scheflen, A.E.: Stream and Structure of Communicational Behavior: Context Analysis of a Psychotherapy Session. Eastern Pennsylvania Psychiatric Institute, Philadelphia (1965)

Scheflen, A.E.: On the structuring of human communication. Am. Behav. Sci. 10 (8), 8–12 (1967). https://doi.org/10.1177/0002764201000803

Schiffrin, D.: Discourse Markers. Studies in Interactional Sociolinguistics, Cambridge University Press, Cambridge (1987). https://doi.org/10.1017/CBO9780511611841

Searle, J.R.: Speech Acts: An Essay in the Philosophy of Language. Cambridge University Press, Cambridge (1969). https://doi.org/10.1017/CBO9781139173438

Searle, J.R.: A classification of illocutionary acts. Lang. Soc. 5 (1), 1–23 (1976). https://doi.org/10.1017/S0047404500006837

Trimboli, C., Walker, M.B.: Switching pauses in cooperative and competitive conversations. J. Exp. Soc. Psychol. 20 (4), 297–311 (1984). https://doi.org/10.1016/0022-1031(84)90027-1

Download references

Acknowledgements

The authors are grateful to the three anonymous reviewers for providing helpful feedback on this paper.

Author information

Authors and affiliations.

Department of Linguistics, Zhejiang University, Hangzhou, China

Da Qi, Chenliang Zhou & Haitao Liu

Center for Linguistics and Applied Linguistics, Guangdong University of Foreign Studies, Guangzhou, China

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Haitao Liu .

Editor information

Editors and affiliations.

Tsinghua University, Beijing, China

Maosong Sun

Harbin Institute of Technology, Harbin, China

Wanxiang Che

Chinese Academy of Sciences, Institute of Computing Technology, Beijing, China

Fudan University, Shanghai, China

Beijing Language and Culture University, Beijing, China

Chinese Academy of Sciences, Institute of Automation, Beijing, China

Appendix A. The Clustering Results of the Frequencies of Discourse Markers in the Speech Acts of Agreement

(See Fig. 3 ).

The clustering results of the frequencies of DMs in the speech acts of Agreement .

Rights and permissions

Reprints and permissions

Copyright information

© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this paper

Cite this paper.

Qi, D., Zhou, C., Liu, H. (2022). Discourse Markers as the Classificatory Factors of Speech Acts. In: Sun, M., et al. Chinese Computational Linguistics. CCL 2022. Lecture Notes in Computer Science(), vol 13603. Springer, Cham. https://doi.org/10.1007/978-3-031-18315-7_1

Download citation

DOI : https://doi.org/10.1007/978-3-031-18315-7_1

Published : 06 October 2022

Publisher Name : Springer, Cham

Print ISBN : 978-3-031-18314-0

Online ISBN : 978-3-031-18315-7

eBook Packages : Computer Science Computer Science (R0)

Share this paper

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Publish with us

Policies and ethics

• Find a journal
• Track your research

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• My Account Login
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 21 December 2023

Towards interpretable speech biomarkers: exploring MFCCs

• Brian Tracey 1 ,
• Dmitri Volfson 1 ,
• James Glass 3 ,
• R’mani Haulcy 3 ,
• Melissa Kostrzebski 2 ,
• Jamie Adams 2 ,
• Tairmae Kangarloo 1 ,
• Amy Brodtmann 4 , 5 ,
• E. Ray Dorsey 2 &
• Adam Vogel 5 , 6  

Scientific Reports volume  13 , Article number:  22787 ( 2023 ) Cite this article

1472 Accesses

1 Altmetric

Metrics details

• Parkinson's disease

While speech biomarkers of disease have attracted increased interest in recent years, a challenge is that features derived from signal processing or machine learning approaches may lack clinical interpretability. As an example, Mel frequency cepstral coefficients (MFCCs) have been identified in several studies as a useful marker of disease, but are regarded as uninterpretable. Here we explore correlations between MFCC coefficients and more interpretable speech biomarkers. In particular we quantify the MFCC2 endpoint, which can be interpreted as a weighted ratio of low- to high-frequency energy, a concept which has been previously linked to disease-induced voice changes. By exploring MFCC2 in several datasets, we show how its sensitivity to disease can be increased by adjusting computation parameters.

Similar content being viewed by others

Relative importance of speech and voice features in the classification of schizophrenia and depression

Mark Berardi, Katharina Brosch, … Maria Dietrich

Speech- and text-based classification of neuropsychiatric conditions in a multidiagnostic setting

Lasse Hansen, Roberta Rocca, … Riccardo Fusaroli

A novel hybrid model integrating MFCC and acoustic parameters for voice disorder detection

Vyom Verma, Anish Benjwal, … Kwok Tai Chui

Introduction

The last decade has seen an increase in the use of speech for health monitoring, with a focus on studies in neurological 1 , 2 and respiratory disease 3 , 4 . This is in part driven by increased ease in recording good-quality data using either smartphones 5 or cloud-based platforms 6 . Analysis of this data has used a mix of interpretable endpoints (prosodic measures related to timing and pitch, etc.) as well as speech parameterizations originally developed for speech recognition. This latter category of parameterizations (which includes MFCCs, or Mel Frequency Cepstral coefficients 1 , RASTA coefficients , and deep learning derived embeddings 7 ) often leads to high performance in classification or regression tasks, but low interpretability. This lack of interpretability makes it difficult to link acoustic features to disease biology and diminishes their utility to clinicians and patients.

MFCCs were originally developed for speech recognition 8 and have found diverse use as voice descriptors, for example in emotions recognition 9 or speech disorder classification 10 . Among MFCC features, the second MFCC coefficient (MFCC2) has been identified as a valuable feature for distinguishing phonation of healthy subjects from people with Parkinson's disease  (PD) 11 , 12 , 13 or other diseases such as major depressive disorder 14 and Alzheimer’s disease 15 .

Speech changes caused by PD are the result of deficits across multiple subsystems of production, including respiration, phonation, articulation, resonance, and prosody 16 . Relevant to MFCCs, respiration is impacted by reduced airflow volume during speech and increased vital capacity percentage per syllable; phonatory deficits result in reduced loudness, and breathy and harsh voice quality; articulatory deficits are characterized by imprecise production of consonants and distorted vowels; and reduced velopharyngeal control manifests in hypernasal resonance. The combination of these changes to speech appears to be reflected in MFCC values, which are thought to model irregular movements in the vocal tract 17 . MFCC features have been shown in multiple studies to significantly differentiate PD subjects from controls, whether used alone 18 or in combination with other features 19 , 20 . Irregular movements can be the result of changes in respiratory support and pressure, altered vocal fold dynamics, or impaired articulator movement-all consequences of PD.

Fronto-temporal dementia (FTD) is another example of a neurodegenerative disease that results in complex speech deficits 21 . There are four variants within the FTD spectrum including the behavioral variant of FTD (bvFTD) and three primary progressive PPA syndromes: nonfluent/agrammatic (nfvPPA), semantic (svPPA), and logopenic (lvPPA). Each present with a distinct communication phenotype, resulting from a combination of cognitive-linguistic and motor impairments. bvFTD has documented speech changes across tasks 2 including reduced rate and accuracy on alternating syllable production tasks. The non-fluent variant yields stark motor speech changes due to apraxia of speech 22 . Errors in vowel and consonant production are common in nfvPPA and lvPPA, but the underlying mechanisms leading to these changes are different. Imprecise production of consonants and vowel distortion in nfvPPA are thought to be the result of impaired motor planning (apraxia) and in some cases a concomitant motor planning deficit (dysarthria). Deficits in sound accuracy in lvPPA are considered a consequence of underlying phonological representation and retrieval, and therefore related to language function. The semantic variant of PPA is largely characterized by word finding deficits and is less associated with frank motor speech impairments.

The discrete speech phenotypes in PD and FTD offer an opportunity to explore the differential impact of disease on commonly used and potentially useful objective markers of speech. As described, MFCCs provide information on vocal tract dynamics, which change based on pathology and manifestations of the disease. The underlying mechanisms driving speech profiles in PD and FTD are different. These contrasts may help explain why we see different MFCC values across diseases and can contribute to our understanding of the metric itself. Data on MFCCs in disease may also build a stronger evidence base for their use in clinical trials and for monitoring disease in the future.

With this motivation, we explored MFCC features (and MFCC2 in particular) in several datasets in PD, frontotemporal dementia (FTD), and healthy speakers. We demonstrate that a) by tuning the MFCC2 calculation to include more high frequencies, we can affect its performance, and b) MFCC2 appears to depend strongly on sex but not age. Finally, we explore correlations between MFCCs (including higher-order MFCCs) and more interpretable voice descriptors.

MFCC computation and interpretation

MFCC2 can be interpreted as a weighted ratio of low- to high-frequency energy, as outlined in the following paragraph. This relationship has been previously noted in the literature 23 although many papers use MFCC features as black-box features. The Discussion reviews existing literature which links low-to-high frequency energy ratios to voice pathology in PD as well as other diseases.

Figure  1 A briefly summarizes the MFCC calculation. The input signal is first transformed to create a spectrogram. Mel frequency filters are then applied to resample the frequency axis in a manner that mimics the roughly logarithmic pitch sensitivity of human hearing, with finer resolution at lower frequencies and coarser resolution at high frequencies. The Mel-filtered data are then log-transformed and processed with a cepstral transform, which amounts to multiplying the Mel spectra by a series of cosine terms. As shown in Fig,  1 A), MFCC1 is a constant feature capturing overall energy, MFCC2 is a half-cycle of cosine, etc. The MFCC coefficients are computed by multiplying the log-transformed Mel spectra by the cosine terms and then summing across frequency.

Figure  1 B shows the cosine term associated with MFCC2, remapped from Mel frequency back to actual frequency in Hz. This figure indicates that MFCC2 is adding a weighted sum of low frequency log(energy) and subtracting off a weighted sum of high frequency log(energy). As \(log(a)-log(b) = log(a/b)\) , MFCC2 can be interpreted as a form of low-to-high frequency energy ratio, with the lowest and highest frequencies contributing most strongly due to the weighting applied.

Figure  1 B also shows two MFCC2 curves. MFCC2 computation requires the user to specify the maximum frequency used in calculation. While 8 kHz is a common upper limit, we show below that increasing the maximum frequency can be beneficial. Figure  1 C shows how low and high frequency spectra can be differentially affected by additive phonation-related noise, as will be discussed in detail below.

Overview of MFCC calculation; ( A ) shows a schematic of the MFCC computation; ( B ) shows the MFCC2 cosine term mapped to frequency in Hz, for maximum frequency values of 8 and 12 kHz; ( C ) shows a ‘cartoon’ view illustrating how increasing aspiration or other noise can impact affect the overall spectrum at high frequencies.

We examined sustained phonation recordings from three datasets, with patient characteristics shown in Table  1 . All three datasets contain control participants. The University of Melbourne/Monash dataset also includes participants with FTD (32 behavioral variant , 11 semantic variant, and 12 nonfluent variant), while the WatchPD dataset includes participants with PD. The PD participants were recruited soon after diagnosis so have fewer years of disease than the FTD participants. Each subject listed in Table 1 provided a single phonation recording. For the Melbourne dataset, a small number of participants made more than one clinic visit, so Tables 2 , 3 include only the first visit for each subject.

Figure  2 shows the averaged acoustic spectra (showing mean and 95th percentile confidence limits for the mean) for the Melbourne and WatchPD datasets, comparing controls to participants with neurological disease (FTD or PD) (called “cases” below). Because spectral characteristics vary by sex, plots are shown separately for males and females. In general, these plots indicate that FTD/PD participants have higher acoustic power at high frequencies as compared to controls. This is seemingly more evident in men, as well as in the FTD participants, who have greater years of disease duration.

Mean spectra (with 95th percentile confidence limits) for Melbourne and WatchPD datasets. Note the generally higher values above 4 kHz for non-healthy participants.

We next computed various low-to-high ratios for our datasets. MFCC2 was computed with the standard 8 kHz maximum frequency, as well as MFCC2 with a 12 kHz maximum frequency. Following Hillenbrand and Houde 24 , we compare energy above and below 4 kHz using an Energy Ratio metric; while they computed a high-to-low ratio, we instead compute a low-to-high ratio for easier comparison to MFCC2 (see Methods). Table  2 lists AUC (area under the curve) values for these metrics from ROC curves for Melbourne and WatchPD datasets (no ROC curves are shown for CLAC as the database only contained control participants). MFCC2 with 12 kHz maximum frequency has the best average AUC, with the energy ratio has the lowest. However, confidence intervals are overlapping, so it is not possible to conclude any particular metric is statistically superior. Descriptive statistics for these metrics are shown in Table  3 . Note that in all cases, control participants have higher mean values for all three metrics.

Next, we analyzed higher-order MFCC coefficients to understand how they map onto more interpretable features. Thus, Fig.  3 shows computed Spearman correlations between MFCC features (mean value and standard deviation across each vocalization) and more interpretable speech features, such as spectral contrast in various frequency ranges 25 , spectral flatness 25 , signal intensity metrics, pitch metrics, and several measures of voice quality (cepstral peak prominence, jitter, shimmer) 26 , 27 , for the Melbourne dataset. The x-axis shows the derived MFCC metrics; the y-axis only includes metrics which showed at least low correlation (>0.3 absolute value) correlation with at least one MFCC metric. Note that multiple features correlate to MFCC2, and that measures of variability (for example standard deviations of spectral contrasts or signal intensity) correlate to MFCC standard deviation metrics.

Pearson correlation coefficients between MFCC features and non-MFCC features (Melbourne data), for non-MFCC features that have at least minimal (> 0.3 absolute value) correlation with at least one MFCC feature. Spectral contrast features are denoted ’Contrast’ with associated frequency ranges; SD denotes standard deviation.

Statistical results

We first explored whether MFCC2 and related metrics were dependent on sex, age and dataset, as well as diagnosis. We modeled each MFCC2 endpoint as a linear combination of factors using the regression function lm in R version 4.2.2. A model selection process using the Akaike Information Criterion (AIC) led to selection of a model which includes gender, dataset, and age. Because dataset and diagnosis are confounded, this analysis was done for control participants only. For MFCC2 with fmax = 12 kHz, there was a highly significant effect of gender (males were higher, p <0.001) and a significant effect of dataset (CLAC values were lower, p <0.05) with no significant effect of age. The corresponding boxplots for MFCC2 with fmax = 12 kHz are shown in Fig.  4 for the different datasets, by gender and diagnosis. MFCC2 values are higher in men, reflecting the increased low-frequency content in these speakers (and are lower in disease, reflecting the increased high-frequency noise seen in Fig.  2 ). For MFCC2 with fmax = 8 kHz (not plotted), these findings were repeated, but also there were also significant effects of age (values decreased with a small slope of roughly 1 point per decade, p<0.05) and also Melbourne values were significantly higher than WatchPD values. Within FTD subjects, we performed ANOVA analysis (after verifying normality assumptions were met) and found no significant differences between FTD subtypes on MFCC2 metrics, though it is important to note our sample sizes are small. Moderate to good correlations were found in each of the three datasets between MFCC2 and the Energy Ratio metric described above (Pearson correlations are 0.82 in the WatchPD data, 0.60 in the Melbourne data, and 0.74 in CLAC).

MFCC2 (fmax = 12 kHz) characteristics, showing dependence on gender and diagnosis. Detailed analysis found in the Statistical Analysis section.

In this section we first discuss findings for the MFCC2 feature, followed by observations about higher-order MFCCs.

The acoustic spectra in Fig.  2 show that (especially in the FTD cohort) differences between cases and controls are small at lower frequencies but are noticeable above roughly 4 kHz. As the MFCC2 can be interpreted as a low-to-high energy ratio, the metric appears to be exploiting this spectral difference to discriminate the presence of disease.

There is a well-established literature that links low-to-high energy ratio differences to voice distortion. Breathy voice can be characterized in the low-frequency range via increased amplitude of the first harmonic, as the glottal waveform becomes more rounded due to non-simultaneous closure along the length of the vocal cords 24 . More relevant to MFCC2, high-frequency energy also increases due to the presence of turbulent airflow associated with breathy voice. Hillenbrand and Houde 24 note, “.. the presence of aspiration noise, which is stronger in the mid and high frequencies than in the lows, can result in a voice signal that is richer in high-frequency energy than nonbreathy signals.” As illustrated in a ‘cartoon’ view in Fig.  1 C, aspiration noise, generated by turbulent flow, tends to fall off less rapidly with frequency than voice. As aspiration noise increases, it primarily impacts the high-frequency part of the overall (voice plus aspiration noise) spectrum but remains negligible compared to voice at the low frequencies. Hillenbrand and Houde thus proposed a high-to-low (H/L) ratio, comparing average energy above 4 kHz to average energy below 4 kHz, to capture breathy voice aspiration noise. Given the high correlation between MFCC2 and this high-low ratio (see Fig.  4 b), one also would expect that MFCC2 would capture breathiness and aspiration noise.

Links between disease and low-high frequency ratio have been established, especially in PD (we are not aware of similar studies in FTD). As reviewed by Ma et al. 28 , multiple studies using laryngoscopic and stroboscopic data have showed that PD is often associated with incomplete glottal closure, asymmetric vocal fold closure, and slower glottal opening, with many of these changes likely due in increased muscle rigidity or poor motor control. Air leakage due to incomplete glottal closure has acoustic effects including increased breathiness and reduced harmonics-to-noise ratio 29 , as additional high frequency noise is present. Incomplete glottal closure can also lead to reduced loudness (hypophonia) due to lower pressure and reduced breath support. The acoustic impacts of incomplete glottal closure are reported to include increased jitter and reduced HNR, also linked to breathiness 16 , 28 .

It is important to note that changes in low-to-high energy ratio differences are not disease-specific. Thus, similar ratios splitting high and low frequencies appear sensitive to changes in speech and voice resulting from Huntington’s disease 30 , illicit drug use 31 , congestion 32 , and fatigue 33 .

Our data shows a clear impact of gender on MFCC2. In healthy participants, MFCC2 values are higher in men than women in all cohorts (Table  3 ), presumably reflecting the fact that male voices are skewed to lower frequencies.

Our results also show that technical issues and computation settings may affect the utility of MFCC2 as a clinical marker. Figure  2 suggests that differences between cases and controls become larger at higher frequencies. The highest upper bound we considered for MFCC computations in this work was 12 kHz. In principle, further increasing the upper range might improve separability between groups. However, manual examination of our data shows occasional high-frequency noise of undetermined source at frequencies above roughly 12-15 kHz. Thus, we limited the upper frequency range to 12 kHz in our calculation. This upper limit is based on engineering judgement, not extensive data exploration. If phonations were recorded in a very quiet environment using high quality hardware 34 , it might be beneficial to include higher frequencies.

Because MFCC2 values depend on the frequency limits used in computation, it is important that MFCC parameter settings should be reported along with findings, to better allow comparisons between different studies. Widely used MFCC implementations such as librosa 25 default to using half the sampling rate as the upper limit, which means that MFCC values could easily vary depending on recording settings (and may include very high-frequency noise, as noted above).

While our focus here is on MFCC2, we also explored higher-order MFCCs. Just as MFCC2 is interpreted as a low-to-high energy comparison, it would be possible to interpret MFCC3 as a mid-frequency to (low+high)-frequency ratio, MFCC4 as a ratio of two frequency bands to two slightly higher frequency bands, etc. (see the cosine shapes in Fig.  1 a), with higher-order MFCCs sampling faster variation across the frequency spectrum. However, the physical / biological meaning of such ratios becomes increasingly unclear as the MFCC coefficient number increases.

We calculated alternate and potentially more intuitive metrics for capturing structure of the acoustic spectra; thus we computed spectral contrast 25 (which measures spectral amplitude peak-to-trough within selected frequency bands), spectral flatness (which measures overall peakiness of the spectrum, with high values representing flat spectra and low values representing tonal-dominate spectra), as well more standard speech features (described in Methods; see also 26 ). Figure 3 shows how the full set of MFCC coefficients correlates to these features, after dropping features with minimal correlation to MFCCs. Several observed correlations are expected; for example, average signal intensities as measured by mean values of MFCC1 and RMS signal intensity are highly correlated. Also, decreased stability of voice intensity is positively correlated with MFCC SDs, as is increased SD of spectral contrast features. More interestingly, increasing voice clarity (as captured by mean values of Cepstral Peak Prominence, or CPP) is negatively correlated to MFCC SDs, suggesting clearer voices also have more stable spectral structure over the phonation. Instability in voice clarity (measured by the SD in cepstral peak prominence or harmonicity) correlates with increased MFCC SD. Instability of the spectral structure as captured by MFCC SD and spectral contrast SD appears correlated; spectral contrast SD potentially is more interpretable in that the frequency bands contributing to instability are identified.

Figure  3 shows several interesting correlations to MFCC2 ( mfcc_mn_2 ). MFCC2 decreases (moves in the direction of pathology) when frequency variability increases, as measured by jitter, pitch SD, and pitch semitone range. MFCC2 also decreases when amplitude variability increases (as measured by shimmer or intensity variability). MFCC2 is higher when the spectrum is more dominated by tonal components (increasing harmonicity or CPP, or decreasing spectral flatness).

Limitations

While we characterized MFCC across several datasets, several of these datasets were relatively small, which limits conclusions that can be drawn. For example, larger datasets would be helpful to characterize potential sex-specific changes in MFCC2 with disease. Table  2 shows that MFCC2 better discriminates Parkinson’s in men (higher AUC). It has been previously established that PD differentially affects male and female voices 1 , 7 . A possible physical explanation is that male voices containing relatively less mid-to-high frequency energy would be more affected by addition of aspiration noise in this frequency range (see Fig.  1 C). However, the same pattern is not observed in FTD participants,which may be related to the small size of the FTD cohort. The small size of our FTD cohort also impacts our ability to perform differential analysis of subtypes. In both cases the uncertainty in the AUCs (seen in the confidence bounds) argues for repeating these analyses in additional datasets.

We also observed differences between datasets (for example greater variability in the normative CLAC dataset) which may be impacted by technical aspects of the recording. The datasets use different recording setups; Melbourne is a mixture of in-lab recordings and at-home (smartphone) recordings, WatchPD consists of in-clinic recordings made with an iPhone (note that only data from the WatchPD baseline visit are analyzed here; the complete WatchPD dataset also includes at-home recordings), and CLAC was recorded via internet browsers using Amazon Mechanical Turk. This means CLAC data are subject to data compression artifacts that may be variable depending on browser, service provider, etc. Manual review uncovered CLAC recordings in which the sustained phonation was distorted after the first second or so, perhaps because noise cancellation algorithms incorrectly identify the sustained phonation as a form of background noise (this issue impacts sustained phonation more than regular speech, as algorithms presumably target hum-like signals). These variations in browser processing may explain the noticeably higher standard deviations seen in metrics computed from CLAC (Table  3 ). This suggests that the acoustic quality of internet-acquired data be carefully reviewed, and that ideally browser settings be controlled to disable processing algorithms, especially when subtle acoustic features of speech are being analyzed.

Given the difficulty in interpreting higher-order MFCCs, interpretable alternative metrics for capturing spectral structure would be of value. As a first step, we performed an initial study of spectral contrast, spectral flatness, and other features, showing correlations to MFCC coefficients (Fig.  3 ). However, further exploration of alternative metrics would be of value.

We analyzed recordings of sustained vowel phonation (“aaah”) from baseline clinic visits in the WatchPD study 5 , and from data collected at the University of Melbourne (consisting of healthy elderly controls 35 and FTD participants 2 ). We also utilize the public-domain CLAC dataset 6 of normative speakers, collected using Amazon Mechanical Turk.

In each dataset, recordings were first automatically segmented using custom Python code to identify the vowel phonation. Processing first detected voiced regions of speech using the voicing/pitch detection from Parselmouth 27 . The initial 0.75 s of voiced data were discarded to remove transient effects, and the next 2.5 s were retained for analysis. If no segment was detected, the recording was not analyzed (and is not included in Tables above). The segmented waveform was then processed using Librosa 25 to compute MFCC coefficients, using 133 Hz as a lower bound and either 8 or 12 kHz as upper bounds. Both mean and standard deviation MFCC features were computed across all frames in each phonation. In addition, the acoustic spectra were computed using the scipy-signal implementation of the Welch periodogram method, using 20 ms, 50% overlapped Hanning windows, as plotted in Fig.  2 . These (linear) spectra \(P_{xx}(f)\) were used to compute the Energy Ratio metric:

where \(f_{max}=12,000\) Hz. Note that whereas Hillenbrand and Houde 24 formed a ratio of high-to-low energy, we compute low-to-high for easier comparison with MFCC2.

Additional (non-MFCC) metrics were computed from the segmented phonation recordings. The Parselmouth-Praat interface 27 was used to compute several speech clarity metrics, including cepstral peak prominence (CPP), harmonicity, jitter (‘localabs’ variant) and shimmer (‘local dB’ variant). The same package was used to compute pitch metrics (mean pitch, standard deviation of pitch, and semitone range). The librosa package was used to explore spectral features, to explore spectral descriptors that are more interpretable than higher-order MFCCs; thus spectral contrast was computed (default settings) as well as spectral contrast in octave bands, starting at 100 Hz (so, 100–200 Hz, 200–400 Hz, etc. up to 3200–6400 Hz).

Statistical analysis was performed in R version 4.1.0. AUC analysis was performed using the pROC package (version 1.18.0) which uses bootstrapping to estimate confidence intervals. For AUCs shown in Table 3 , MFCC2 was the only feature; thus no classifier is required, as MFCC2 can be used as a scalar test statistic. AUC results in Table 3 were generated by sweeping the threshold values across the range of MFCC2.

Data availability

Raw audio for the CLAC dataset is available at https://groups.csail.mit.edu/sls/downloads/clac/. The extracted features for the CLAC dataset are available at https://github.com/brianhtracey/mfcc2_related. Extracted features for other datasets may be available upon reasonable request.

Code availability

Core python code for MFCC2 feature extraction is available at https://github.com/brianhtracey/mfcc2_related.

Tsanas, A., Little, M. A., McSharry, P. E., Spielman, J. & Ramig, L. O. Novel speech signal processing algorithms for high-accuracy classification of Parkinson’s disease. IEEE Trans. Biomed. Eng. 59 , 1264–1271 (2012).

Article   PubMed   Google Scholar  

Vogel, A. P. et al. Motor speech signature of behavioral variant frontotemporal dementia: Refining the phenotype. Neurology 89 , 837–844 (2017).

Quatieri, T. F., Talkar, T. & Palmer, J. S. A framework for biomarkers of covid-19 based on coordination of speech-production subsystems. IEEE Open J. Eng. Med. Biol. 1 , 203–206 (2020).

Tracey, B. et al. Voice biomarkers of recovery from acute respiratory illness. IEEE J. Biomed. Health Inform. 26 (6), 2787–2795 (2021).

Article   Google Scholar  

Cedarbaum, J. M. et al. Enabling efficient use of digital health technologies to support parkinson’s disease drug development through precompetitive collaboration. In American Society for Clinical Pharmacology & Therapeutics (ASCPT) Meeting (2019).

Haulcy, R. & Glass, J. CLAC: A Speech Corpus of Healthy English Speakers. In Proceedings of the Interspeech 2021 , 2966–2970, https://doi.org/10.21437/Interspeech.2021-1810 (2021).

Jeancolas, L. et al. X-vectors: New quantitative biomarkers for early Parkinson’s disease detection from speech. Front. Neuroinform. 15 , 578369 (2021).

Article   PubMed   PubMed Central   Google Scholar  

Davis, S. & Mermelstein, P. Comparison of parametric representations for monosyllabic word recognition in continuously spoken sentences. IEEE Trans. Acoust. Speech Signal Process. 28 , 357–366 (1980).

Kathiresan, T. & Dellwo, V. Cepstral derivatives in mfccs for emotion recognition. In 2019 IEEE 4th International Conference on Signal and Image Processing (ICSIP) , 56–60 (IEEE, 2019).

Ai, O. C., Hariharan, M., Yaacob, S. & Chee, L. S. Classification of speech dysfluencies with mfcc and lpcc features. Expert Syst. Appl. 39 , 2157–2165 (2012).

Lipsmeier, F. et al. Evaluation of smartphone-based testing to generate exploratory outcome measures in a phase 1 Parkinson’s disease clinical trial. Mov. Disord. 33 , 1287–1297 (2018).

Kapoor, T. & Sharma, R. Parkinson’s disease diagnosis using mel-frequency cepstral coefficients and vector quantization. Int. J. Comput. Appl. 14 , 43–46 (2011).

Google Scholar  

Benba, A., Jilbab, A. & Hammouch, A. Detecting patients with Parkinson’s disease using mel frequency cepstral coefficients and support vector machines. Int. J. Electr. Eng. Inform. 7 , 297 (2015).

Taguchi, T. et al. Major depressive disorder discrimination using vocal acoustic features. J. Affect. Disord. 225 , 214–220 (2018).

Al-Hameed, S., Benaissa, M. & Christensen, H. Simple and robust audio-based detection of biomarkers for Alzheimer’s disease. In 7th Workshop on Speech and Language Processing for Assistive Technologies (SLPAT) , 32–36 (2016).

Magee, M., Copland, D. & Vogel, A. P. Motor speech and non-motor language endophenotypes of Parkinson’s disease. Expert Rev. Neurother. 19 , 1191–1200 (2019).

Article   CAS   PubMed   Google Scholar  

Godino-Llorente, J. I., Gomez-Vilda, P. & Blanco-Velasco, M. Dimensionality reduction of a pathological voice quality assessment system based on gaussian mixture models and short-term cepstral parameters. IEEE Trans. Biomed. Eng. 53 , 1943–1953 (2006).

Benba, A., Jilbab, A., Hammouch, A. & Sandabad, S. Voiceprints analysis using mfcc and svm for detecting patients with Parkinson’s disease. In 2015 International Conference on Electrical and Information Technologies (ICEIT) , 300–304 (IEEE, 2015).

Tsanas, A., Little, M. A., Fox, C. & Ramig, L. O. Objective automatic assessment of rehabilitative speech treatment in Parkinson’s disease. IEEE Trans. Neural Syst. Rehabil. Eng. 22 , 181–190 (2013).

Hawi, S. et al. Automatic Parkinson’s disease detection based on the combination of long-term acoustic features and mel frequency cepstral coefficients (mfcc). Biomed. Signal Process. Control 78 , 104013 (2022).

Poole, M. L., Brodtmann, A., Darby, D. & Vogel, A. P. Motor speech phenotypes of frontotemporal dementia, primary progressive aphasia, and progressive apraxia of speech. J. Speech Lang. Hear. Res. 60 , 897–911 (2017).

Ogar, J. M., Dronkers, N. F., Brambati, S. M., Miller, B. L. & Gorno-Tempini, M. L. Progressive nonfluent aphasia and its characteristic motor speech deficits. Alzheimer Dis. Assoc. Disord. 21 , S23–S30 (2007).

Hlavnička, J. et al. Automated analysis of connected speech reveals early biomarkers of Parkinson’s disease in patients with rapid eye movement sleep behaviour disorder. Sci. Rep. 7 , 12 (2017).

Article   ADS   PubMed   PubMed Central   Google Scholar  

Hillenbrand, J. & Houde, R. A. Acoustic correlates of breathy vocal quality: Dysphonic voices and continuous speech. J. Speech Lang. Hear. Res. 39 , 311–321 (1996).

Article   CAS   Google Scholar  

McFee, B. et al . librosa 0.5.0, https://doi.org/10.5281/zenodo.293021 (2017).

Schultz, B. G. & Vogel, A. P. A tutorial review on clinical acoustic markers in speech science. J. Speech Lang. Hear. Res. 65 , 3239–3263 (2022).

Jadoul, Y., Thompson, B. & de Boer, B. Introducing parselmouth: A python interface to praat. J. Phon. 71 , 1–15. https://doi.org/10.1016/j.wocn.2018.07.001 (2018).

Ma, A., Lau, K. K. & Thyagarajan, D. Voice changes in parkinson’s disease: What are they telling us?. J. Clin. Neurosci. 72 , 1–7 (2020).

Bhuta, T., Patrick, L. & Garnett, J. D. Perceptual evaluation of voice quality and its correlation with acoustic measurements. J. Voice 18 , 299–304 (2004).

Vogel, A. P., Shirbin, C., Churchyard, A. J. & Stout, J. C. Speech acoustic markers of early stage and prodromal Huntington’s disease: A marker of disease onset?. Neuropsychologia 50 , 3273–3278 (2012).

Vogel, A. P. et al. Adults with a history of recreational cannabis use have altered speech production. Drug Alcohol Depend. 227 , 108963 (2021).

Lee, G.-S., Yang, C. C., Wang, C.-P. & Kuo, T. B. Effect of nasal decongestion on voice spectrum of a nasal consonant-vowel. J. Voice 19 , 71–77 (2005).

Vogel, A. P., Fletcher, J. & Maruff, P. Acoustic analysis of the effects of sustained wakefulness on speech. J. Acoust. Soc. Am. 128 , 3747–3756 (2010).

Article   ADS   PubMed   Google Scholar  

Vogel, A. P. & Reece, H. Recording speech: Methods and formats. In Manual of Clinical Phonetics , 217–227 (Routledge, 2021).

Schultz, B. G., Rojas, S., St John, M., Kefalianos, E. & Vogel, A. P. A cross-sectional study of perceptual and acoustic voice characteristics in healthy aging. J. Voice (2021).

Download references

Acknowledgements

This research was supported by Millennium Pharmaceuticals, Inc. (a subsidiary of Takeda Pharmaceuticals). Funding for the WatchPD study was contributed by Biogen, Takeda, and the members of the Critical Path for Parkinson’s Consortium 3DT Initiative, Stage 2. The authors acknowledge the members and advisors of the CPP 3DT initiative and the FDA for their insightful feedback to the WATCH-PD study design and execution. The authors gratefully acknowledge WatchPD participants who joined WatchPD study participant forums, which helped clarify the need for more interpretable voice features. We also acknowledge Dr. Sandra Rojas as contributor of healthy control data from Australia.

Author information

Authors and affiliations.

Takeda Pharamaceuticals, Data Science Institute, Cambridge, MA, 02142, USA

Brian Tracey, Dmitri Volfson & Tairmae Kangarloo

Center for Health + Technology (CHeT), University of Rochester Medical Center, Rochester, NY, USA

Melissa Kostrzebski, Jamie Adams & E. Ray Dorsey

Massachusetts Institute of Technology, CSAIL, Cambridge, MA, 02139, USA

James Glass & R’mani Haulcy

Monash University, Melbourne, VIC, Australia

Amy Brodtmann

University of Melbourne, Parkville, VIC, 3010, Australia

Amy Brodtmann & Adam Vogel

Redenlab Inc, Melbourne, VIC, 3010, Australia

You can also search for this author in PubMed   Google Scholar

Contributions

B.T., D.V., and J.G. conceived the experiment(s), M.K., J.A., R.D., R.H., J.G., T.K., and A.B. conducted the experiment(s) and contributed participants, B.T. and D.V. analysed the results. All authors reviewed the manuscript.

Corresponding author

Correspondence to Brian Tracey .

Ethics declarations

Competing interests.

B. Tracey, D. Volfson and T. Kangarloo are full-time employees of and own stock in Takeda Pharmaceuticals. M. Kostrzebski holds stock in Apple, Inc. J. Adams has received compensation for consulting services from VisualDx and the Huntington Study Group; and research support from Biogen, Biosensics, Huntington Study Group, Michael J. Fox Foundation, National Institutes of Health/National Institute of Neurological Disorders and Stroke, NeuroNext Network, and Safra Foundation. E. Dorsey has received compensation for consulting services from Abbott, Abbvie, Acadia, Acorda, Bial-Biotech Investments, Inc., Biogen, Boehringer Ingelheim, California Pacific Medical Center, Caraway Therapeutics, Curasen Therapeutics, Denali Therapeutics, Eli Lilly, Genentech/Roche, Grand Rounds, Huntington Study Group, Informa Pharma Consulting, Karger Publications, LifeSciences Consultants, MCM Education, Mediflix, Medopad, Medrhythms, Merck, Michael J. Fox Foundation, NACCME, Neurocrine, NeuroDerm, NIH, Novartis, Origent Data Sciences, Otsuka, Physician’s Education Resource, Praxis, PRIME Education, Roach, Brown, McCarthy and Gruber, Sanofi, Seminal Healthcare, Spark, Springer Healthcare, Sunovion Pharma, Theravance, Voyager and WebMD; research support from Biosensics, Burroughs Wellcome Fund, CuraSen, Greater Rochester Health Foundation, Huntington Study Group, Michael J. Fox Foundation, National Institutes of Health, Patient-Centered Outcomes Research Institute, Pfizer, PhotoPharmics, Safra Foundation, and Wave Life Sciences; editorial services for Karger Publications; stock in Included Health and in Mediflix, and ownership interests in SemCap. J. Glass and R. Haulcy have received research funding from Takeda Pharmaceuticals. A.Vogel is Chief Science Officer of Redenlab P/L who undertake speech biomarker research services. A. Brodtmann has no competing financial interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Tracey, B., Volfson, D., Glass, J. et al. Towards interpretable speech biomarkers: exploring MFCCs. Sci Rep 13 , 22787 (2023). https://doi.org/10.1038/s41598-023-49352-2

Download citation

Received : 14 January 2023

Accepted : 07 December 2023

Published : 21 December 2023

DOI : https://doi.org/10.1038/s41598-023-49352-2

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing: Translational Research newsletter — top stories in biotechnology, drug discovery and pharma.

BioVoix™, Speech analysis for health indicators

Analyzing micro speech markers ™ in your voice, an advanced biomarker.

‘Speech Markers ™ ’ is a revolutionary leading-edge research startup working in the biomedical field. Our ground-breaking, patent-pending technology platform, BioVoix™, uses the biomarkers in human speech samples to determine, measure and predict various health and physical parameters, including the presence of abnormalities, diseases, pregnancies and other conditions. The diagnostics are performed by analyzing the complex relationship between speech characteristics and various human physiological attributes such as hormone levels, organ dimensions, cardiovascular, respiratory and other conditions.

At present, Speech Markers is in the process of entering collaboration with hospitals, clinics and medical experts to conduct further research and validation of our implementation, leading eventually to clinical trials and an initial launch of our solution in India.

Health speaks, we understand

What our speechmarkers™ reveal, hormone levels.

Hormone level estimates for FSH, LH, estrogen, progesterone, hCG

Organ Dimensions

Estimated dimensions and volume for ovaries, uterus, kidney, and many more...

Cardiac Health

Overall cardiovascular health, blood pressure and heart rate

Hormone levels, pelvic region dimensions, fetus dimensions and more...

Neurological Health

Neurological disorders, short term & long term memory issues

Fertility Health

Ovulation days, menstruation stages & phases, hormonal changes, and more...

Respiratory Health

Respiratory system

Women's Health

Breast function abnormality, menstruation stages & phases, hormone levels, fertility and more...

Diabetic Health

Blood sugar level

Speech Markers™ Easy realtime access to health indicators

Speech Markers™ is developing technology to identify micro markers in your speech that change with your health condition. To receive a health report, you simply speak the words that our mobile app shows you on the screen.

Use your Device

We have developed a mobile application (BioVoix ™) to help you send us your speech sample. Our team analyzes the SpeechMarkers™ in your speech samples and delivers your health report.

BioVoix is the first of its kind that uses your speech as input to determine your health. You speak the letters, words and sentences. That's it. Really!

Underlying Science

The micro changes in your voice are due to subtle changes in your physiological conditions. We have conducted extensive study to corelate speech characteristics with health conditions.

Confidential

We respect your right to privacy. Your data remains confidential. Strong security underlies all our technology processes.

Over three decades of research

Our founding team has over three decades of extensive research experience. We have analyzed thousands of speech samples and have published our research.

Drop us a line

We are based in Pune, India

We are always looking for bright and curious minds to join our team

We would love to hear from you. Write to us. [email protected]

Deep Spanish

Free spanish lessons and readings

Your cart is currently empty!

Lesson 2 – Spanish Speech Markers (Marcadores del Discurso)

What are “los marcadores del discurso”.

The Spanish speech markers (“Los marcadores del discurso”) are linguistic elements that help to structure and organize a text or speech. They serve as connectors or links between sentences, providing cohesion and coherence to the discourse. In Spanish, these markers can be words or phrases that indicate the relationship between the ideas expressed, making it easier for the reader or listener to follow the flow of the text.

Let’s start with the connectors

Connectors (los conectores)

These are the markers that link one part of the speech with a previous one. Their function is to maintain logical relationships or dependencies between the ideas conveyed by different statements.

They can be

• María y Juan fueron al cine.
• No me gusta la comida picante. Además , soy alérgico a los camarones.
• Me gusta leer libros de ciencia ficción. También disfruto de las películas de ese género.
• Estudia mucho para sus exámenes. Asimismo , participa en actividades extracurriculares.
• No me gusta el café ni el té.

Consecutivos

• No tenía dinero, entonces no pudo comprar el libro.
• Estaba lloviendo, por lo tanto , decidimos quedarnos en casa.
• No había luz en la casa, así que encendimos velas.
• No estudió para el examen, en consecuencia , reprobó.
• El tráfico estaba terrible, de modo que llegamos tarde al evento.

Contraargumentativos

• Me gusta el cine, pero no quiero ir hoy.
• Estaba cansado, sin embargo , siguió trabajando.
• Aunque tenía miedo, decidió enfrentar la situación.
• Sabía que era difícil, no obstante , se esforzó al máximo.
• A él le gusta el fútbol, en cambio , a ella le gusta el baloncesto.

Justificativos

• No fui a la fiesta porque estaba enfermo.
• No puedo ir al cine contigo ya que tengo mucho trabajo.
• No salimos a caminar puesto que estaba lloviendo.
• El vuelo se retrasó debido a mal tiempo.
• La carretera estaba cerrada a causa de un accidente.

Condicionales

• Si estudias, aprobarás el examen.
• Lleva un paraguas en caso de que llueva.
• No podrás entrar a menos que tengas una invitación.
• Puedes usar mi coche siempre que lo devuelvas antes de las 10 pm.
• Te ayudaré a condición de que me lo pidas.

Information structurers

• These speech markers serve to signal the beginning, continuity, or closure of an idea within a text or speech. They help to guide the reader or listener through the development of the argument or narrative.

They can be “de”

• Para empezar , quiero decir que estoy de acuerdo con la propuesta.
• En primer lugar , debemos analizar la situación actual.
• Antes que nada , agradezco la oportunidad de hablar ante ustedes.
• En segundo lugar , es importante considerar las posibles soluciones.
• La empresa ha tenido éxito en el pasado. Además , cuenta con un equipo altamente capacitado.
• El proyecto es costoso. Por otro lado , tiene el potencial de generar ingresos significativos.
• En resumen , la propuesta tiene ventajas y desventajas, pero vale la pena considerarla.
• Por último , quiero agradecer a todos por su atención y participación.
• En conclusión , creo que debemos seguir adelante con el proyecto.

Comentadores (Commentary)

• El proyecto es ambicioso, es decir , requerirá de mucho esfuerzo y dedicación.
• La situación es complicada; en otras palabras , no hay una solución fácil.
• A mi parecer , la mejor opción es seguir adelante con la propuesta.
• Desde mi punto de vista , debemos ser cautelosos al tomar una decisión.
• En efecto , los datos muestran que la estrategia ha sido exitosa.

Digresores (Digressions)

• A propósito , ¿has hablado con Juan sobre el tema?
• Por cierto , me encontré con María ayer y me comentó sobre los cambios en la empresa.
• A todo esto , ¿qué pasó con el informe que ibas a presentar?
• Volviendo al tema , creo que debemos centrarnos en las prioridades principales.
• En cuanto a la propuesta, creo que debemos analizarla detenidamente antes de tomar una decisión.

Reformulating speech markers

are used to rephrase, clarify or summarize information within a text or speech. They help to ensure that the intended message is conveyed accurately and effectively. We will explore four main categories of reformulating discourse markers: explicativos, rectificativos, de distanciamiento, and recapitulativos.

Explicativos (Explanatory)

• La situación es complicada; o sea , no hay una solución fácil.
• No estoy de acuerdo con la propuesta. Me explico , creo que hay otras opciones más viables.
• No me gusta el diseño. Aclaro , no es que esté mal hecho, sino que prefiero otro estilo.
• El proceso es lento y tedioso; en otras palabras , llevará tiempo completarlo.

Rectificativos (Rectifying)

• El proyecto es costoso, mejor dicho , es una inversión importante.
• La reunión es mañana, no hoy. Rectifico , me equivoqué en la fecha.
• En realidad , la situación no es tan grave como parece.
• De hecho , los datos demuestran que la estrategia ha sido exitosa.
• El problema no es la falta de recursos, más bien es la mala administración de los mismos.

De Distanciamiento (Distancing)

• La propuesta tiene ventajas. Sin embargo , también presenta desventajas importantes.
• Aunque entiendo tu punto de vista, no estoy de acuerdo con él.
• El equipo ha trabajado duro. No obstante , los resultados no son los esperados.
• La idea es interesante, pero creo que necesita más desarrollo.
• A pesar de los contratiempos, seguimos adelante con el proyecto.

Recapitulativos (Recapitulating)

• En síntesis , el proyecto presenta oportunidades de crecimiento y desarrollo.
• Para concluir , creo que debemos seguir adelante con el proyecto.
• En pocas palabras , la situación es complicada pero no imposible de resolver.
• En suma , los beneficios superan los riesgos y debemos continuar con la propuesta.

Argumentative operators

“Los operadores argumentativos,” are essential tools for constructing logical and persuasive arguments in both written and spoken Spanish. They help to establish relationships between ideas and guide the reader or listener through the argument. We will examine two main categories of argumentative operators: gradativos and no gradativos.

Gradativos (Gradual)

• En primer lugar , debemos analizar los datos disponibles.
• Además , es importante tener en cuenta las opiniones de los expertos.
• También debemos considerar las posibles consecuencias a largo plazo.
• Asimismo , es crucial evaluar los costos y beneficios.
• Por último , debemos tomar una decisión basada en la información recopilada.

No Gradativos (Non-gradual)

No gradativos are argumentative operators that do not indicate a gradual progression or hierarchy within an argument. Instead, they help to establish relationships between ideas, such as cause and effect, contrast, or concession. Some common no gradativos include:

Causales (Causal)

• No podemos continuar con el proyecto porque no tenemos suficientes recursos.
• Puesto que no hay consenso, debemos buscar otra solución.
• No es posible avanzar ya que no hemos recibido la aprobación necesaria.

Concesivos (Concessive)

• Aunque hay argumentos en contra, creo que debemos seguir adelante.
• La propuesta tiene sus desventajas, sin embargo , sus beneficios son mayores.
• No obstante las dificultades, el equipo logró cumplir con los objetivos.

Contrastivos (Contrastive)

• La idea es interesante, pero necesita más desarrollo.
• La opción A es más costosa, en cambio , la opción B es más económica.
• La empresa X ha tenido éxito, mientras que la empresa Y ha enfrentado dificultades.

Informal Discourse Markers

• ¿Vamos al cine? – Pues sí, me parece buena idea. (Shall we go to the movies? – Well, yes, that sounds like a good idea.)
• ¿Qué te parece la propuesta? – Pues … no sé, tengo que pensarlo. (What do you think about the proposal? – Well… I don’t know, I have to think about it.)
• No puedo ir a la fiesta, pues tengo que estudiar para un examen. (I can’t go to the party, as I have to study for an exam.)
• Ya entiendo lo que quieres decir. (Now I understand what you’re saying.)
• Ya veo que tienes razón. (I see that you’re right.)
• Ya terminé el trabajo. (I’ve already finished the work.)
• Hombre , ¡no me esperaba verte aquí! (Man, I didn’t expect to see you here!)
• Hombre , no estoy de acuerdo con esa idea. (Man, I don’t agree with that idea.)
• Hombre , tienes que ver esta película. (Man, you have to see this movie.)
• Mujer , ¡cuánto tiempo sin verte! (Girl, it’s been so long since I’ve seen you!)
• Mujer , creo que te estás equivocando. (Girl, I think you’re mistaken.)
• Mujer , este restaurante es increíble. (Girl, this restaurant is amazing.)
• Hija , ¿estás bien? Te ves cansada. (Sweetheart, are you okay? You look tired.)
• Hijo , ¡no sabía que venías a visitarme! (Son, I didn’t know you were coming to visit me!)
• Hija , no te preocupes tanto por lo que piensen los demás. (Sweetheart, don’t worry so much about what others think.)

Speech-Language Pathology and Audiology Programs Rank Among Best Graduate Programs

U.S. News and World Report has ranked University of Wisconsin-Madison Speech-Language Pathology and University of Wisconsin AuD Consortium as “Best Programs”, placing them among the elite graduate school programs in the nation for 2024.

• Speech-Language Pathology (2)
• Audiology (16)

The national survey rankings reflect the peer assessments of academic quality, but other considerations involving location, environment, strength of different fields, cost after tuition and financial aid, and job placement are also very important.

Related: UW–Madison graduate programs ranked highly by U.S. News

Wrentham educators and supporters block speech ban

With more than 2,000 emails sent to the Wrentham School Committee via a campaign organized by the Wrentham Educators Association to oppose a proposed policy to restrict speech on social policies in Wrentham schools, the board appears ready to drop the plan when it meets Tuesday.

The proposed policy sought to prohibit, “advocacy through the use of pamphlets, stickers, pins, buttons, insignias, flags, banners, posters, signs, photographs, or other similar materials.” Social policies are vaguely defined as “topics of national, state, or local interest, over which the public is deeply divided and are often intensely personal or important to adherents, which are the source of conflicting opinions on the grounds of what is perceived as morally correct or incorrect, or which are the subject to intense partisan advocacy or debate.”

Educators maintained that the proposed policy was vague and would allow micromanagement of school staff in ways that could stifle free speech and silence support for marginalized communities. 

Members of WEA and their supporters will attend the School Committee meeting scheduled for 7 p.m. Tuesday in the Gibbons Gym at the Delaney School.

The WEA issued the following statement:

“WEA members were pleased to learn that the proposed policy change banning ‘advocacy’ is set to be withdrawn. Supporting our students and creating an environment where everyone feels welcome and ready to learn requires an array of tools and strategies. Educators saw this proposed policy as vague and potentially restrictive in ways that would have undermined our ability to exercise professional judgment. Wrentham educators respect the diversity of perspectives and backgrounds in our schools and in our community. We value the trust placed in us to treat everyone fairly and with dignity.”



Subscribe to MTA headlines MTA Today

We've detected unusual activity from your computer network

To continue, please click the box below to let us know you're not a robot.

Why did this happen?

Please make sure your browser supports JavaScript and cookies and that you are not blocking them from loading. For more information you can review our Terms of Service and Cookie Policy .

For inquiries related to this message please contact our support team and provide the reference ID below.